tristan/gptel
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
gptel/gptel-curl.el

356 lines
16 KiB
EmacsLisp
Raw Normal View History

gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
;;; gptel-curl.el --- Curl support for GPTel -*- lexical-binding: t; -*-
;; Copyright (C) 2023 Karthik Chikmagalur
;; Author: Karthik Chikmagalur;; <karthikchikmagalur@gmail.com>
;; Keywords: convenience
gptel-curl: Autoload gptel-curl-get-response gptel-curl.el (gptel-curl-get-response): Rename from `gptel--curl-get-response' and autoload it to ease its use in `gptel-send'. Remove Version header identifying gptel-curl as a separate package and make it require `gptel' instead.
2023-03-08 19:17:14 -08:00
;; SPDX-License-Identifier: GPL-3.0-or-later
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
;; This program is free software; you can redistribute it and/or modify
;; it under the terms of the GNU General Public License as published by
;; the Free Software Foundation, either version 3 of the License, or
;; (at your option) any later version.
;; This program is distributed in the hope that it will be useful,
;; but WITHOUT ANY WARRANTY; without even the implied warranty of
;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
;; GNU General Public License for more details.
;; You should have received a copy of the GNU General Public License
;; along with this program. If not, see <https://www.gnu.org/licenses/>.
;;; Commentary:
;; Curl support for GPTel. Utility functions.
;;; Code:
gptel-curl: Autoload gptel-curl-get-response gptel-curl.el (gptel-curl-get-response): Rename from `gptel--curl-get-response' and autoload it to ease its use in `gptel-send'. Remove Version header identifying gptel-curl as a separate package and make it require `gptel' instead.
2023-03-08 19:17:14 -08:00
(require 'gptel)
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
(eval-when-compile
gptel: curl proxy support (#69) * gptel.el (gptel-proxy): Support a proxy when interacting with openai endpoint. In many organizations the openai api can only be accessed via proxy. This is easily supported by curl. gptel-curl.el (gptel-curl--get-args): tidy up `gptel-curl--get-args'. --------- Co-authored-by: PalaceChan <XXX>
2023-06-05 23:17:55 -05:00
(require 'cl-lib)
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
(require 'subr-x))
(require 'map)
(require 'json)
gptel-curl: Rename functions for linting gptel-curl.el (gptel-curl--process-alist, gptel-curl--get-args, gptel--curl-sentinel, gptel-curl--parse-response): Rename internal functions and variables to use the `gptel-curl--` prefix instead of `gptel--curl-`.
2023-03-08 19:23:17 -08:00
(defvar gptel-curl--process-alist nil
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
"Alist of active GPTel curl requests.")
gptel-curl: Rename functions for linting gptel-curl.el (gptel-curl--process-alist, gptel-curl--get-args, gptel--curl-sentinel, gptel-curl--parse-response): Rename internal functions and variables to use the `gptel-curl--` prefix instead of `gptel--curl-`.
2023-03-08 19:23:17 -08:00
(defun gptel-curl--get-args (prompts token)
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
"Produce list of arguments for calling Curl.
PROMPTS is the data to send, TOKEN is a unique identifier."
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(let* ((url (gptel-backend-url gptel-backend))
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
(data (encode-coding-string
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(json-encode (gptel--request-data gptel-backend prompts))
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
'utf-8))
(headers
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(append '(("Content-Type" . "application/json"))
(when-let ((backend-header (gptel-backend-header gptel-backend)))
(if (functionp backend-header)
(funcall backend-header)
backend-header)))))
gptel: curl proxy support (#69) * gptel.el (gptel-proxy): Support a proxy when interacting with openai endpoint. In many organizations the openai api can only be accessed via proxy. This is easily supported by curl. gptel-curl.el (gptel-curl--get-args): tidy up `gptel-curl--get-args'. --------- Co-authored-by: PalaceChan <XXX>
2023-06-05 23:17:55 -05:00
(append
(list "--location" "--silent" "--compressed" "--disable"
(format "-X%s" "POST")
(format "-w(%s . %%{size_header})" token)
(format "-m%s" 60)
"-D-"
(format "-d%s" data))
(when (not (string-empty-p gptel-proxy))
(list "--proxy" gptel-proxy
"--proxy-negotiate"
"--proxy-user" ":"))
(cl-loop for (key . val) in headers
collect (format "-H%s: %s" key val))
(list url))))
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
gptel: org support for streaming WIP * gptel.el (gptel--convert-playback-markdown->org): New converter for markdown->org that works on text chunks while maintaining the parse state until the text stream is finished. * gptel-curl.el (gptel--insert-response-stream, gptel-curl-get-response): When using `gptel-playback' and requesting ChatGPT's responses in org-mode, run the above converter on the received response. This works by storing the converter and associated state as a closure in the async info plist that is supplied along with the response, and running it repeatedly on each chunk of text in the response stream before it is inserted into the buffer. FIXME: Note that `gptel-response-filter-functions' is currently ignored if using `gptel-stream'.
2023-04-04 22:19:37 -07:00
;;TODO: The :transformer argument here is an alternate implementation of
;;`gptel-response-filter-functions'. The two need to be unified.
gptel-curl: Autoload gptel-curl-get-response gptel-curl.el (gptel-curl-get-response): Rename from `gptel--curl-get-response' and autoload it to ease its use in `gptel-send'. Remove Version header identifying gptel-curl as a separate package and make it require `gptel' instead.
2023-03-08 19:17:14 -08:00
;;;###autoload
gptel: More flexible callbacks * gptel.el (gptel--url-get-response, gptel-api-key-from-auth-source): `gptel--url-get-response' accepts a callback argument that can be used to do something besides inserting the response into the current buffer. * gptel-curl.el (gptel-curl--sentinel, gptel-curl-get-response): `gptel-curl--sentinel' now accepts a callback argument that can be used to do something besides inserting the response into the current buffer. These changes are in preparation for more specific functionality, like showing the response as a message, or replacing the prompt with the response etc.
2023-03-31 19:37:25 -07:00
(defun gptel-curl-get-response (info &optional callback)
gptel: Remove aio dependency * gptel.el (gptel-send, gptel--insert-response, gptel--url-get-response): Remove aio dependency, turn aio-defuns into regular functions. This requires splitting `gptel-send' into "before response" and "after response" functions, but the ability to debug the code fully is worth the inconvenience. The new "after response" function is `gptel--insert-response'. * gptel-curl.el (gptel-curl--sentinel, gptel-curl-get-response): Turn aio-defuns into regular functions.
2023-03-17 23:46:30 -07:00
"Retrieve response to prompt in INFO.
INFO is a plist with the following keys:
- :prompt (the prompt being sent)
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
- :buffer (the gptel buffer)
- :position (marker at which to insert the response).
gptel: More flexible callbacks * gptel.el (gptel--url-get-response, gptel-api-key-from-auth-source): `gptel--url-get-response' accepts a callback argument that can be used to do something besides inserting the response into the current buffer. * gptel-curl.el (gptel-curl--sentinel, gptel-curl-get-response): `gptel-curl--sentinel' now accepts a callback argument that can be used to do something besides inserting the response into the current buffer. These changes are in preparation for more specific functionality, like showing the response as a message, or replacing the prompt with the response etc.
2023-03-31 19:37:25 -07:00
Call CALLBACK with the response and INFO afterwards. If omitted
the response is inserted into the current buffer after point."
gptel-curl: Fix current-buffer for Curl requests * gptel-curl.el (gptel-curl-get-response): Set buffer-local model parameters in the correct (i.e. gptel) buffer, not in Curl's process buffer. This fixes #43.
2023-04-12 22:10:54 -07:00
(let* ((token (md5 (format "%s%s%s%s"
(random) (emacs-pid) (user-full-name)
(recent-keys))))
(args (gptel-curl--get-args (plist-get info :prompt) token))
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(stream (and gptel-stream (gptel-backend-stream gptel-backend)))
gptel-curl: Fix current-buffer for Curl requests * gptel-curl.el (gptel-curl-get-response): Set buffer-local model parameters in the correct (i.e. gptel) buffer, not in Curl's process buffer. This fixes #43.
2023-04-12 22:10:54 -07:00
(process (apply #'start-process "gptel-curl"
(generate-new-buffer "*gptel-curl*") "curl" args)))
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(when gptel--debug
(message "%S" args))
gptel-curl: Fix current-buffer for Curl requests * gptel-curl.el (gptel-curl-get-response): Set buffer-local model parameters in the correct (i.e. gptel) buffer, not in Curl's process buffer. This fixes #43.
2023-04-12 22:10:54 -07:00
(with-current-buffer (process-buffer process)
gptel: Remove aio dependency * gptel.el (gptel-send, gptel--insert-response, gptel--url-get-response): Remove aio dependency, turn aio-defuns into regular functions. This requires splitting `gptel-send' into "before response" and "after response" functions, but the ability to debug the code fully is worth the inconvenience. The new "after response" function is `gptel--insert-response'. * gptel-curl.el (gptel-curl--sentinel, gptel-curl-get-response): Turn aio-defuns into regular functions.
2023-03-17 23:46:30 -07:00
(set-process-query-on-exit-flag process nil)
(setf (alist-get process gptel-curl--process-alist)
gptel: More flexible callbacks * gptel.el (gptel--url-get-response, gptel-api-key-from-auth-source): `gptel--url-get-response' accepts a callback argument that can be used to do something besides inserting the response into the current buffer. * gptel-curl.el (gptel-curl--sentinel, gptel-curl-get-response): `gptel-curl--sentinel' now accepts a callback argument that can be used to do something besides inserting the response into the current buffer. These changes are in preparation for more specific functionality, like showing the response as a message, or replacing the prompt with the response etc.
2023-03-31 19:37:25 -07:00
(nconc (list :token token
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
;; FIXME `aref' breaks `cl-struct' abstraction boundary
;; FIXME `cl--generic-method' is an internal `cl-struct'
:parser (cl--generic-method-function
(if stream
(cl-find-method
'gptel-curl--parse-stream nil
(list
(aref (buffer-local-value
'gptel-backend (plist-get info :buffer))
0) t))
(cl-find-method
'gptel--parse-response nil
(list
(aref (buffer-local-value
'gptel-backend (plist-get info :buffer))
0) t t))))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
:callback (or callback
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(if stream
gptel: Rename stream-related vars and defuns * gptel.el (gptel--convert-playback-markdown->org, gptel--request-data, gptel-playback): - `gptel-playback' is now obsolete, use `gptel-stream' instead. - `gptel--convert-playback-markdown->org': rename to `gptel--stream-convert-markdown->org' * gptel-curl.el (gptel-curl--filter, gptel-curl-get-response): - `gptel--insert-response-stream': rename to `gptel-curl--stream-insert-response' - `gptel-curl--cleanup-stream': rename to `gptel-curl--stream-cleanup' - `gptel-curl--filter': rename to `gptel-curl--stream-filter'
2023-04-06 16:35:51 -07:00
#'gptel-curl--stream-insert-response
gptel: org support for streaming WIP * gptel.el (gptel--convert-playback-markdown->org): New converter for markdown->org that works on text chunks while maintaining the parse state until the text stream is finished. * gptel-curl.el (gptel--insert-response-stream, gptel-curl-get-response): When using `gptel-playback' and requesting ChatGPT's responses in org-mode, run the above converter on the received response. This works by storing the converter and associated state as a closure in the async info plist that is supplied along with the response, and running it repeatedly on each chunk of text in the response stream before it is inserted into the buffer. FIXME: Note that `gptel-response-filter-functions' is currently ignored if using `gptel-stream'.
2023-04-04 22:19:37 -07:00
#'gptel--insert-response))
gptel-curl: Only convert to Org in Org buffers * gptel-curl.el (gptel-curl-get-response): Don't convert response into org-mode unless the buffer from which the request originated is in org-mode. This makes `gptel-default-mode' less binding, and only used when creating a new chat session with `gptel'. Also, gptel should now do the right thing depending on whether the current buffer is in text, Markdown or Org modes.
2023-07-21 13:32:07 -07:00
:transformer (when (eq (buffer-local-value
'major-mode
(plist-get info :buffer))
'org-mode)
gptel: Rename stream-related vars and defuns * gptel.el (gptel--convert-playback-markdown->org, gptel--request-data, gptel-playback): - `gptel-playback' is now obsolete, use `gptel-stream' instead. - `gptel--convert-playback-markdown->org': rename to `gptel--stream-convert-markdown->org' * gptel-curl.el (gptel-curl--filter, gptel-curl-get-response): - `gptel--insert-response-stream': rename to `gptel-curl--stream-insert-response' - `gptel-curl--cleanup-stream': rename to `gptel-curl--stream-cleanup' - `gptel-curl--filter': rename to `gptel-curl--stream-filter'
2023-04-06 16:35:51 -07:00
(gptel--stream-convert-markdown->org)))
gptel: More flexible callbacks * gptel.el (gptel--url-get-response, gptel-api-key-from-auth-source): `gptel--url-get-response' accepts a callback argument that can be used to do something besides inserting the response into the current buffer. * gptel-curl.el (gptel-curl--sentinel, gptel-curl-get-response): `gptel-curl--sentinel' now accepts a callback argument that can be used to do something besides inserting the response into the current buffer. These changes are in preparation for more specific functionality, like showing the response as a message, or replacing the prompt with the response etc.
2023-03-31 19:37:25 -07:00
info))
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(if stream
gptel: Rename stream-related vars and defuns * gptel.el (gptel--convert-playback-markdown->org, gptel--request-data, gptel-playback): - `gptel-playback' is now obsolete, use `gptel-stream' instead. - `gptel--convert-playback-markdown->org': rename to `gptel--stream-convert-markdown->org' * gptel-curl.el (gptel-curl--filter, gptel-curl-get-response): - `gptel--insert-response-stream': rename to `gptel-curl--stream-insert-response' - `gptel-curl--cleanup-stream': rename to `gptel-curl--stream-cleanup' - `gptel-curl--filter': rename to `gptel-curl--stream-filter'
2023-04-06 16:35:51 -07:00
(progn (set-process-sentinel process #'gptel-curl--stream-cleanup)
(set-process-filter process #'gptel-curl--stream-filter))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(set-process-sentinel process #'gptel-curl--sentinel)))))
gptel-curl: Add gptel-abort * gptel-curl.el (gptel-abort): Command to cancel any active gptel Curl request in the current buffer.
2023-05-19 22:26:38 -07:00
(defun gptel-abort (buf)
"Stop any active gptel process associated with the current buffer."
(interactive (list (current-buffer)))
(unless gptel-use-curl
(user-error "Cannot stop a `url-retrieve' request!"))
(if-let* ((proc-attrs
(cl-find-if
(lambda (proc-list)
(eq (plist-get (cdr proc-list) :buffer) buf))
gptel-curl--process-alist))
(proc (car proc-attrs)))
(progn
(setf (alist-get proc gptel-curl--process-alist nil 'remove) nil)
(set-process-sentinel proc #'ignore)
(delete-process proc)
(kill-buffer (process-buffer proc))
(with-current-buffer buf
(when gptel-mode (gptel--update-header-line " Ready" 'success)))
(message "Stopped gptel request in buffer %S" (buffer-name buf)))
(message "No gptel request associated with buffer %S" (buffer-name buf))))
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
;; TODO: Separate user-messaging from this function
gptel, gptel-curl: Fix byte-compile warnings gptel.el (gptel--request-data): Also use :json-false to encode nil in the http request.
2023-06-23 16:43:16 -07:00
(defun gptel-curl--stream-cleanup (process _status)
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
"Process sentinel for GPTel curl requests.
gptel, gptel-curl: Fix byte-compile warnings gptel.el (gptel--request-data): Also use :json-false to encode nil in the http request.
2023-06-23 16:43:16 -07:00
PROCESS and _STATUS are process parameters."
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(let ((proc-buf (process-buffer process)))
(when gptel--debug
(with-current-buffer proc-buf
(clone-buffer "*gptel-error*" 'show)))
(let* ((info (alist-get process gptel-curl--process-alist))
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(gptel-buffer (plist-get info :buffer))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(tracking-marker (plist-get info :tracking-marker))
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(start-marker (plist-get info :position))
gptel-curl: Better error handling when streaming * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-cleanup): When streaming responses, move the error handling from the Curl process filter to the cleanup sentinel. This simplifies the filter code a fair bit.
2023-04-08 09:57:38 -07:00
(http-status (plist-get info :http-status))
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(http-msg (plist-get info :status)))
gptel-curl: Better error handling when streaming * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-cleanup): When streaming responses, move the error handling from the Curl process filter to the cleanup sentinel. This simplifies the filter code a fair bit.
2023-04-08 09:57:38 -07:00
(if (equal http-status "200")
gptel: Handle read-only gptel-buffers * gptel.el (gptel--insert-response): * gptel-transient.el (gptel--suffix-send): * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup): Handle read-only gptel buffers by redirecting the output to a new buffer (that pops up automatically). To track this, - the `:position' argument of the INFO plist, which is a marker, is moved to the new output buffer. - the `:buffer' argument of the INFO plist is unmodified, it always points to the buffer that the request originated from.
2023-05-03 13:42:44 -07:00
(progn
;; Finish handling response
(with-current-buffer (marker-buffer start-marker)
(pulse-momentary-highlight-region (+ start-marker 2) tracking-marker)
(when gptel-mode (save-excursion (goto-char tracking-marker)
(insert "\n\n" (gptel-prompt-string)))))
(with-current-buffer gptel-buffer
(when gptel-mode (gptel--update-header-line " Ready" 'success))))
gptel-curl: Better error handling when streaming * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-cleanup): When streaming responses, move the error handling from the Curl process filter to the cleanup sentinel. This simplifies the filter code a fair bit.
2023-04-08 09:57:38 -07:00
;; Or Capture error message
(with-current-buffer proc-buf
(goto-char (point-max))
(search-backward (plist-get info :token))
(backward-char)
(pcase-let* ((`(,_ . ,header-size) (read (current-buffer)))
gptel: Unify error descriptions * gptel.el (gptel--url-parse-response, gptel--insert-response): Use the same error codes/descriptions across url-retrieve/Curl, with and without streaming responses. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--stream-filter, gptel-curl--stream-cleanup): Ditto.
2023-04-08 12:22:36 -07:00
(json-object-type 'plist)
gptel-curl: Better error handling when streaming * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-cleanup): When streaming responses, move the error handling from the Curl process filter to the cleanup sentinel. This simplifies the filter code a fair bit.
2023-04-08 09:57:38 -07:00
(response (progn (goto-char header-size)
(condition-case nil (json-read)
(json-readtable-error 'json-read-error)))))
(cond
((plist-get response :error)
(let* ((error-plist (plist-get response :error))
(error-msg (plist-get error-plist :message))
(error-type (plist-get error-plist :type)))
gptel: Unify error descriptions * gptel.el (gptel--url-parse-response, gptel--insert-response): Use the same error codes/descriptions across url-retrieve/Curl, with and without streaming responses. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--stream-filter, gptel-curl--stream-cleanup): Ditto.
2023-04-08 12:22:36 -07:00
(message "ChatGPT error: (%s) %s" http-msg error-msg)
(setq http-msg (concat "(" http-msg ") " (string-trim error-type)))))
gptel-curl: Better error handling when streaming * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-cleanup): When streaming responses, move the error handling from the Curl process filter to the cleanup sentinel. This simplifies the filter code a fair bit.
2023-04-08 09:57:38 -07:00
((eq response 'json-read-error)
(message "ChatGPT error (%s): Malformed JSON in response." http-msg))
(t (message "ChatGPT error (%s): Could not parse HTTP response." http-msg)))))
(with-current-buffer gptel-buffer
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(when gptel-mode
gptel-curl: Better error handling when streaming * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-cleanup): When streaming responses, move the error handling from the Curl process filter to the cleanup sentinel. This simplifies the filter code a fair bit.
2023-04-08 09:57:38 -07:00
(gptel--update-header-line
gptel-curl: Run post-response-hook in gptel buffer * gptel-curl.el (gptel-curl--stream-cleanup): `gptel-post-response-hook' should run in the buffer that was current when the request was sent. This was not the case for the curl method (with response streaming). Fixed.
2023-05-03 13:47:15 -07:00
(format " Response Error: %s" http-msg) 'error))))
(with-current-buffer gptel-buffer
(run-hooks 'gptel-post-response-hook)))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(setf (alist-get process gptel-curl--process-alist nil 'remove) nil)
(kill-buffer proc-buf)))
gptel: Rename :insert-marker * gptel.el (gptel--url-get-response, gptel--insert-response, gptel-send): Rename the :insert-marker keyword in the async info plist to :start-marker. * gptel-curl.el (gptel--insert-response-stream, gptel-curl--cleanup-stream, gptel-curl-get-response): Ditto.
2023-04-06 16:18:37 -07:00
(defun gptel-curl--stream-insert-response (response info)
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
"Insert streaming RESPONSE from ChatGPT into the gptel buffer.
INFO is a mutable plist containing information relevant to this buffer.
See `gptel--url-get-response' for details."
gptel, gptel-curl: Fix byte-compile warnings gptel.el (gptel--request-data): Also use :json-false to encode nil in the http request.
2023-06-23 16:43:16 -07:00
(let ((start-marker (plist-get info :position))
gptel: org support for streaming WIP * gptel.el (gptel--convert-playback-markdown->org): New converter for markdown->org that works on text chunks while maintaining the parse state until the text stream is finished. * gptel-curl.el (gptel--insert-response-stream, gptel-curl-get-response): When using `gptel-playback' and requesting ChatGPT's responses in org-mode, run the above converter on the received response. This works by storing the converter and associated state as a closure in the async info plist that is supplied along with the response, and running it repeatedly on each chunk of text in the response stream before it is inserted into the buffer. FIXME: Note that `gptel-response-filter-functions' is currently ignored if using `gptel-stream'.
2023-04-04 22:19:37 -07:00
(tracking-marker (plist-get info :tracking-marker))
(transformer (plist-get info :transformer)))
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(when response
gptel: Handle read-only gptel-buffers * gptel.el (gptel--insert-response): * gptel-transient.el (gptel--suffix-send): * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup): Handle read-only gptel buffers by redirecting the output to a new buffer (that pops up automatically). To track this, - the `:position' argument of the INFO plist, which is a marker, is moved to the new output buffer. - the `:buffer' argument of the INFO plist is unmodified, it always points to the buffer that the request originated from.
2023-05-03 13:42:44 -07:00
(with-current-buffer (marker-buffer start-marker)
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(save-excursion
(unless tracking-marker
(gptel--update-header-line " Typing..." 'success)
(goto-char start-marker)
gptel: Handle read-only gptel-buffers * gptel.el (gptel--insert-response): * gptel-transient.el (gptel--suffix-send): * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup): Handle read-only gptel buffers by redirecting the output to a new buffer (that pops up automatically). To track this, - the `:position' argument of the INFO plist, which is a marker, is moved to the new output buffer. - the `:buffer' argument of the INFO plist is unmodified, it always points to the buffer that the request originated from.
2023-05-03 13:42:44 -07:00
(unless (or (bobp) (plist-get info :in-place))
gptel: Add streaming, in-place options to gptel-request * gptel.el (gptel--insert-response, gptel-request): - Add an in-place key to gptel-request. When true, the default callbacks will not delimit the API responses with newlines. - Add a strea option to gptel-request. Only works with the default filter/stream-insert callback, so it's marked as for internal use for now. * gptel-curl.el (gptel-curl--stream-insert-response): Ditto.
2023-04-09 03:29:48 -07:00
(insert "\n\n"))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(setq tracking-marker (set-marker (make-marker) (point)))
(set-marker-insertion-type tracking-marker t)
(plist-put info :tracking-marker tracking-marker))
gptel: org support for streaming WIP * gptel.el (gptel--convert-playback-markdown->org): New converter for markdown->org that works on text chunks while maintaining the parse state until the text stream is finished. * gptel-curl.el (gptel--insert-response-stream, gptel-curl-get-response): When using `gptel-playback' and requesting ChatGPT's responses in org-mode, run the above converter on the received response. This works by storing the converter and associated state as a closure in the async info plist that is supplied along with the response, and running it repeatedly on each chunk of text in the response stream before it is inserted into the buffer. FIXME: Note that `gptel-response-filter-functions' is currently ignored if using `gptel-stream'.
2023-04-04 22:19:37 -07:00
(when transformer
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(setq response (funcall transformer response)))
gptel: org support for streaming WIP * gptel.el (gptel--convert-playback-markdown->org): New converter for markdown->org that works on text chunks while maintaining the parse state until the text stream is finished. * gptel-curl.el (gptel--insert-response-stream, gptel-curl-get-response): When using `gptel-playback' and requesting ChatGPT's responses in org-mode, run the above converter on the received response. This works by storing the converter and associated state as a closure in the async info plist that is supplied along with the response, and running it repeatedly on each chunk of text in the response stream before it is inserted into the buffer. FIXME: Note that `gptel-response-filter-functions' is currently ignored if using `gptel-stream'.
2023-04-04 22:19:37 -07:00
gptel, gptel-curl: Make the gptel text-property non-sticky gptel.el (gptel--insert-response): gptel-curl.el (gptel-curl--stream-insert-response): Make the `gptel' text-property rear-nonsticky so typing after it is recognized as part of the user prompt.
2023-10-28 19:34:54 -07:00
(add-text-properties
0 (length response) '(gptel response rear-nonsticky t)
response)
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(goto-char tracking-marker)
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(insert response))))))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
gptel: Rename stream-related vars and defuns * gptel.el (gptel--convert-playback-markdown->org, gptel--request-data, gptel-playback): - `gptel-playback' is now obsolete, use `gptel-stream' instead. - `gptel--convert-playback-markdown->org': rename to `gptel--stream-convert-markdown->org' * gptel-curl.el (gptel-curl--filter, gptel-curl-get-response): - `gptel--insert-response-stream': rename to `gptel-curl--stream-insert-response' - `gptel-curl--cleanup-stream': rename to `gptel-curl--stream-cleanup' - `gptel-curl--filter': rename to `gptel-curl--stream-filter'
2023-04-06 16:35:51 -07:00
(defun gptel-curl--stream-filter (process output)
gptel, gptel-curl: Fix byte-compile warnings gptel.el (gptel--request-data): Also use :json-false to encode nil in the http request.
2023-06-23 16:43:16 -07:00
(let* ((proc-info (alist-get process gptel-curl--process-alist)))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(with-current-buffer (process-buffer process)
;; Insert output
(save-excursion
(goto-char (process-mark process))
(insert output)
(set-marker (process-mark process) (point)))
;; Find HTTP status
(unless (plist-get proc-info :http-status)
(save-excursion
(goto-char (point-min))
(when-let* (((not (= (line-end-position) (point-max))))
(http-msg (buffer-substring (line-beginning-position)
(line-end-position)))
(http-status
(save-match-data
(and (string-match "HTTP/[.0-9]+ +\\([0-9]+\\)" http-msg)
(match-string 1 http-msg)))))
(plist-put proc-info :http-status http-status)
gptel: Handle read-only gptel-buffers * gptel.el (gptel--insert-response): * gptel-transient.el (gptel--suffix-send): * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup): Handle read-only gptel buffers by redirecting the output to a new buffer (that pops up automatically). To track this, - the `:position' argument of the INFO plist, which is a marker, is moved to the new output buffer. - the `:buffer' argument of the INFO plist is unmodified, it always points to the buffer that the request originated from.
2023-05-03 13:42:44 -07:00
(plist-put proc-info :status (string-trim http-msg))))
;; Handle read-only gptel buffer
(when (with-current-buffer (plist-get proc-info :buffer)
(or buffer-read-only
(get-char-property (plist-get proc-info :position) 'read-only)))
(message "Buffer is read only, displaying reply in buffer \"*ChatGPT response*\"")
(display-buffer
(with-current-buffer (get-buffer-create "*ChatGPT response*")
(goto-char (point-max))
(move-marker (plist-get proc-info :position) (point) (current-buffer))
(current-buffer))
'((display-buffer-reuse-window
display-buffer-pop-up-window)
gptel: Add pre-response-hook * gptel.el (gptel--insert-response, gptel-pre-response-hook): New user option `gptel-pre-response-hook' that runs before the response is inserted into the buffer. This can be used to prepare the buffer in some user-specified way for the response. * gptel-curl.el (gptel-curl--stream-filter): Run `gptel-pre-response-hook' before inserting streaming responses.
2023-07-25 15:28:51 -07:00
(reusable-frames . visible))))
;; Run pre-response hook
(when (and (equal (plist-get proc-info :http-status) "200")
gptel-pre-response-hook)
(with-current-buffer (marker-buffer (plist-get proc-info :position))
(run-hooks 'gptel-pre-response-hook))))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(when-let ((http-msg (plist-get proc-info :status))
gptel-curl: process filter for streaming support * gptel.el (gptel--request-data): Request a streaming message if `gptel-stream' is non-nil. * gptel-curl.el (gptel-curl-get-response, gptel-curl--cleanup-stream, gptel-curl--filter): Add a process filter and sentinel for Curl to stream ChatGPT's response into Emacs in real-time.
2023-04-01 02:36:10 -07:00
(http-status (plist-get proc-info :http-status)))
;; Find data chunk(s) and run callback
gptel-curl: Better error handling when streaming * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-cleanup): When streaming responses, move the error handling from the Curl process filter to the cleanup sentinel. This simplifies the filter code a fair bit.
2023-04-08 09:57:38 -07:00
(when (equal http-status "200")
(funcall (or (plist-get proc-info :callback)
#'gptel-curl--stream-insert-response)
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(funcall (plist-get proc-info :parser) nil proc-info)
gptel-curl: Better error handling when streaming * gptel-curl.el (gptel-curl--stream-filter, gptel-curl--stream-cleanup): When streaming responses, move the error handling from the Curl process filter to the cleanup sentinel. This simplifies the filter code a fair bit.
2023-04-08 09:57:38 -07:00
proc-info))))))
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(cl-defgeneric gptel-curl--parse-stream (backend proc-info)
"Stream parser for gptel-curl.
Implementations of this function run as part of the process
filter for the active query, and return partial responses from
the LLM.
BACKEND is the LLM backend in use.
PROC-INFO is a plist with process information and other context.
See `gptel-curl--get-response' for its contents.")
gptel, gptel-curl: Fix byte-compile warnings gptel.el (gptel--request-data): Also use :json-false to encode nil in the http request.
2023-06-23 16:43:16 -07:00
(defun gptel-curl--sentinel (process _status)
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
"Process sentinel for GPTel curl requests.
gptel, gptel-curl: Fix byte-compile warnings gptel.el (gptel--request-data): Also use :json-false to encode nil in the http request.
2023-06-23 16:43:16 -07:00
PROCESS and _STATUS are process parameters."
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
(let ((proc-buf (process-buffer process)))
gptel: Add a debug flag gptel.el (gptel--debug, gptel--url-parse-response): Add a debug flag that shows the http response. Fix json parsing error. gptel-curl.el (gptel-curl--sentinel): Ditto.
2023-03-12 15:40:39 -07:00
(when gptel--debug
(with-current-buffer proc-buf
(clone-buffer "*gptel-error*" 'show)))
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(when-let* (((eq (process-status process) 'exit))
(proc-info (alist-get process gptel-curl--process-alist))
(proc-callback (plist-get proc-info :callback)))
(pcase-let ((`(,response ,http-msg ,error)
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(with-current-buffer proc-buf
(gptel-curl--parse-response proc-info))))
gptel: Simplify response API * gptel.el (gptel--url-parse-response, gptel--url-get-response, gptel--insert-response, gptel-send): - Use shorter keys for passing the info plist, - record errors in the info plist, - separate user messaging from the callback and more. - Make the API more functional (i.e. less imperative) This is in preparation for adding `gptel-request', an API for defining custom commands. Note: The streaming filter and callback are mostly unchanged. Streaming is not planned to be accessible via `gptel-request'. * gptel-curl.el (gptel-curl--parse-response, gptel-curl--sentinel, gptel-curl--stream-filter, gptel-curl--stream-insert-response, gptel-curl--stream-cleanup, gptel-curl-get-response): Ditto.
2023-04-08 16:57:38 -07:00
(plist-put proc-info :status http-msg)
(when error (plist-put proc-info :error error))
(funcall proc-callback response proc-info)))
gptel: Remove aio dependency * gptel.el (gptel-send, gptel--insert-response, gptel--url-get-response): Remove aio dependency, turn aio-defuns into regular functions. This requires splitting `gptel-send' into "before response" and "after response" functions, but the ability to debug the code fully is worth the inconvenience. The new "after response" function is `gptel--insert-response'. * gptel-curl.el (gptel-curl--sentinel, gptel-curl-get-response): Turn aio-defuns into regular functions.
2023-03-17 23:46:30 -07:00
(setf (alist-get process gptel-curl--process-alist nil 'remove) nil)
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
(kill-buffer proc-buf)))
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(defun gptel-curl--parse-response (proc-info)
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
"Parse the buffer BUF with curl's response.
TOKEN is used to disambiguate multiple requests in a single
buffer."
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(let ((token (plist-get proc-info :token))
(parser (plist-get proc-info :parser)))
(goto-char (point-max))
(search-backward token)
(backward-char)
(pcase-let* ((`(,_ . ,header-size) (read (current-buffer))))
(goto-char (point-min))
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
gptel: Add multi-llm support README.org: Update README with new information and a multi-llm demo. gptel.el (gptel-host, gptel--known-backends, gptel--api-key, gptel--create-prompt, gptel--request-data, gptel--parse-buffer, gptel-request, gptel--parse-response, gptel--openai, gptel--debug, gptel--restore-state, gptel, gptel-backend): Integrate multiple LLMs through the introcution of gptel-backends. Each backend is composed of two pieces: 1. An instance of a cl-struct, containing connection, authentication and model information. See the cl-struct `gptel-backend` for details. A separate cl-struct type is defined for each supported backend (OpenAI, Azure, GPT4All and Ollama) that inherits from the generic gptel-backend type. 2. cl-generic implementations of specific tasks, like gathering up and formatting context (previous user queries and LLM responses), parsing responses or responses streams etc. The four tasks currently specialized this way are carried out by `gptel--parse-buffer` and `gptel--request-data` (for constructing the query) and `gptel--parse-response` and `gptel-curl--parse-stream` (for parsing the response). See their implementations for details. Some effort has been made to limit the number of times dispatching is done when reading streaming responses. When a backend is created, it is registered in the collection `gptel--known-backends` and can be accessed by name later, such as from the transient menu. Only one of these backends is active at any time in a buffer, stored in the buffer-local variable `gptel-backend`. Most messaging, authentication etc accounts for the active backend, although there might be some leftovers. When using `gptel-request` or `gptel-send`, the active backend can be changed or let-bound. - Obsolete `gptel-host` - Fix the rear-sticky property when restoring sessions from files. - Document some variables (not user options), like `gptel--debug` gptel-openai.el (gptel-backend, gptel-make-openai, gptel-make-azure, gptel-make-gpt4all): This file (currently always loaded) sets up the generic backend struct and includes constructors for creating OpenAI, GPT4All and Azure backends. They all use the same API so a single set of defgeneric implemenations suffices for all of them. gptel-ollama.el (gptel-make-ollama): This file includes the cl-struct, constructor and requisite defgeneric implementations for Ollama support. gptel-transient.el (gptel-menu, gptel-provider-variable, gptel--infix-provider, gptel-suffix-send): - Provide access to all available LLM backends and models from `gptel-menu`. - Adjust keybindings in gptel-menu: setting the model and query parameters is now bound to two char keybinds, while redirecting input and output is bound to single keys.
2023-10-23 23:40:52 -07:00
(if-let* ((http-msg (string-trim
(buffer-substring (line-beginning-position)
(line-end-position))))
(http-status
(save-match-data
(and (string-match "HTTP/[.0-9]+ +\\([0-9]+\\)" http-msg)
(match-string 1 http-msg))))
(json-object-type 'plist)
(response (progn (goto-char header-size)
(condition-case nil
(json-read)
(json-readtable-error 'json-read-error)))))
(cond
((equal http-status "200")
(list (string-trim
(funcall parser nil response proc-info))
http-msg))
((plist-get response :error)
(let* ((error-plist (plist-get response :error))
(error-msg (plist-get error-plist :message))
(error-type (plist-get error-plist :type)))
(list nil (concat "(" http-msg ") " (string-trim error-type)) error-msg)))
((eq response 'json-read-error)
(list nil (concat "(" http-msg ") Malformed JSON in response.")
"Malformed JSON in response"))
(t (list nil (concat "(" http-msg ") Could not parse HTTP response.")
"Could not parse HTTP response.")))
(list nil (concat "(" http-msg ") Could not parse HTTP response.")
"Could not parse HTTP response.")))))
gptel-curl: Add curl module and playback feature. Conditionally solves #2. gptel.el (gptel-use-curl, gptel-parse-response, gptel--playback, gptel-send, gptel-playback): New user options `gptel-playback', `gptel-use-curl`. The former controls whether the response is played back in chunks, which is done by the function `gptel--playback'. The response returned by `gptel-get-response' and `gptel--curl-get-response' is now a plist with the content and status. gptel-curl.el (gptel--curl-get-args, gptel--curl-get-response, gptel--curl-sentinel): Add support for curl when available. Set it to the default. `url-retrieve' is full of fangs that multibyte you.
2023-03-08 00:52:48 -08:00
(provide 'gptel-curl)
;;; gptel-curl.el ends here
Reference in a new issue Copy permalink