gptel: Add gptel--request-params for extra request options

* gptel.el (gptel--request-params)
(gptel--with-buffer-copy-internal): Add new variable for
specifying global/buffer-local/oneshot request-params.

Important: It is very easy to break gptel with stale/incompatible
request-params.  As such, request-params are best specified as
backend or model properties.  If used, `gptel--request-params' is
best set via a preset, and it's the user's responsibility to unset
this when switching presets.  For this reason it is marked as an
internal variable.  Really, it is only here because of the unique
configuration requirements of John Wiegley.

* gptel-anthropic.el (gptel--request-data): Use it in request
payload.
* gptel-gemini.el (gptel--request-data): Use it in request
payload.
* gptel-ollama.el (gptel--request-data): Use it in request
payload.
* gptel-openai.el (gptel--request-data): Use it in request
payload.
* gptel-kagi.el:  Update TODO about request-params support.

Author: karthink

gptel-anthropic.el

@@ -233,6 +233,7 @@ Mutate state INFO with response metadata."
     ;; Merge request params with model and backend params.
     (gptel--merge-plists
      prompts-plist
+     gptel--request-params
      (gptel-backend-request-params gptel-backend)
      (gptel--model-request-params  gptel-model))))
 

gptel-gemini.el

@@ -151,6 +151,7 @@ list."
     ;; Merge request params with model and backend params.
     (gptel--merge-plists
      prompts-plist
+     gptel--request-params
      (gptel-backend-request-params gptel-backend)
      (gptel--model-request-params  gptel-model))))
 

gptel-kagi.el

@@ -71,7 +71,8 @@
                      (concat "\n\n" (mapconcat #'identity ref-strings "\n")))))
     (concat output references)))
 
-;; TODO: Add model and backend-specific request-params support
+;; TODO: Add support for model/backend-specific request-params and
+;; gptel--request-params
 (cl-defmethod gptel--request-data ((_backend gptel-kagi) prompts)
   "JSON encode PROMPTS for Kagi."
   (pcase-exhaustive (gptel--model-name gptel-model)

gptel-ollama.el

@@ -102,6 +102,7 @@ Store response metadata in state INFO."
            `(:model ,(gptel--model-name gptel-model)
              :messages [,@prompts]
              :stream ,(or gptel-stream :json-false))
+           gptel--request-params
            (gptel-backend-request-params gptel-backend)
            (gptel--model-request-params  gptel-model)))
          ;; the initial options (if any) from request params

gptel-openai.el

@@ -312,6 +312,7 @@ Mutate state INFO with response metadata."
     ;; Merge request params with model and backend params.
     (gptel--merge-plists
      prompts-plist
+     gptel--request-params
      (gptel-backend-request-params gptel-backend)
      (gptel--model-request-params  gptel-model))))
 

gptel.el

@@ -906,6 +906,14 @@ or
 Each entry has the form (PROCESS . (FSM ABORT-CLOSURE))
 If the ABORT-CLOSURE is called, it must abort the PROCESS.")
 
+(defvar gptel--request-params nil
+  "Extra parameters sent with each gptel request.
+
+These parameters are combined with model-specific and backend-specific
+:request-params before sending a request, which see.  Warning: values
+incompatible with the active backend can break gptel.  Do not use this
+variable unless you know what you're doing!")
+
 
 ;;; Utility functions
 
@@ -1086,7 +1094,7 @@ For BUF, START, END and BODY-THUNK see `gptel--with-buffer-copy'."
                       gptel-mode gptel-track-response gptel-track-media
                       gptel-use-tools gptel-tools gptel-use-curl
                       gptel-use-context gptel--num-messages-to-send
-                      gptel-stream gptel-include-reasoning
+                      gptel-stream gptel-include-reasoning gptel--request-params
                       gptel-temperature gptel-max-tokens gptel-cache))
         (set (make-local-variable sym) (buffer-local-value sym buf)))
       (when (and start end) (insert-buffer-substring buf start end))