[nongnu] elpa/gptel 4273f067e8 271/273: gptel-org: Improve stream converter

[ELPA Syncer]

branch: elpa/gptel
commit 4273f067e8e60e14e797071bd87bd0f617d03996
Author: Karthik Chikmagalur <karthikchikmagalur@gmail.com>
Commit: Karthik Chikmagalur <karthikchikmagalur@gmail.com>

    gptel-org: Improve stream converter
    
    * test/gptel-org-test.el (gptel-org--test-stream-conversion): Add
    a test harness to make comparing markdown to org easy.
    
    * gptel-org.el (gptel--stream-convert-markdown->org): Handle
    headings and strong chars in the converter.  Addresses #296.
---
 gptel-org.el           | 20 +++++++++---
 test/gptel-org-test.el | 89 ++++++++++++++++++++++++++++++--------------------
 2 files changed, 68 insertions(+), 41 deletions(-)

diff --git a/gptel-org.el b/gptel-org.el
index 7f3dc5c72e..5b12564867 100644
--- a/gptel-org.el
+++ b/gptel-org.el
@@ -407,7 +407,7 @@ text stream."
           (when (marker-position start-pt) (goto-char start-pt))
           (when in-src-block (setq ticks ticks-total))
           (save-excursion
-            (while (re-search-forward "`\\|\\*\\{1,2\\}\\|_" nil t)
+            (while (re-search-forward "`\\|\\*\\{1,2\\}\\|_\\|^#+" nil t)
               (pcase (match-string 0)
                 ("`"
                  ;; Count number of consecutive backticks
@@ -439,12 +439,22 @@ text stream."
                     ;; Negative number of ticks or in a src block already,
                     ;; reset ticks
                     (t (setq ticks ticks-total)))))
-                ;; Handle other chars: emphasis, bold and bullet items
+                ;; Handle other chars: heading, emphasis, bold and bullet items
+                ((and (guard (and (not in-src-block) (eq (char-before) ?#))) 
heading)
+                 (if (eobp)
+                     ;; Not enough information about the heading yet
+                     (progn (setq noop-p t) (set-marker start-pt 
(match-beginning 0)))
+                   ;; Convert markdown heading to Org heading
+                   (when (looking-at "[[:space:]]")
+                     (delete-region (line-beginning-position) (point))
+                     (insert (make-string (length heading) ?*)))))
                 ((and "**" (guard (not in-src-block)))
                  (cond
-                  ((looking-at "\\*\\(?:[[:word:]]\\|\s\\)")
-                   (delete-char 1))
-                  ((looking-back "\\(?:[[:word:]]\\|\s\\)\\*\\{2\\}"
+                  ;; TODO Not sure why this branch was needed
+                  ;; ((looking-at "\\*\\(?:[[:word:]]\\|\s\\)") (delete-char 
1))
+
+                  ;; Looking back at "w**" or " **"
+                  ((looking-back "\\(?:[[:word:][:punct:]\n]\\|\s\\)\\*\\{2\\}"
                                  (max (- (point) 3) (point-min)))
                    (delete-char -1))))
                 ((and "*" (guard (not in-src-block)))
diff --git a/test/gptel-org-test.el b/test/gptel-org-test.el
index c7ad3c5d7e..91be6e8499 100644
--- a/test/gptel-org-test.el
+++ b/test/gptel-org-test.el
@@ -26,9 +26,20 @@
           (:success strs)))
       (setq strs (delq nil (nreverse strs))))))
 
-;;; Basic tests for markdown to org conversion
-(let ((string-sequences
-       '(("" "```" "cpp" "
+(defmacro gptel-org--test-stream-conversion (md-list org-str)
+  `(ert-deftest
+       ,(intern (concat "gptel-org--test-stream-conversion-"
+                 (substring (sha1 (prin1-to-string md-list)) 0 10)))
+       ()
+       (let ((func (gptel--stream-convert-markdown->org)))
+        (prog1
+            (should
+             (string= (apply #'concat (mapcar func ,md-list))
+              ,org-str))
+          (setq gptel-post-response-functions nil)))))
+
+(gptel-org--test-stream-conversion
+ '("" "```" "cpp" "
 " "#include" " <" "cstdio" ">
 
 " "int" " main" "()" " {
@@ -36,7 +47,18 @@
 " "   " " return" " " "0" ";
 " "}
 " "```")
-        ("" "Here" " is" " a" " simple" " C" "++" " program" " that" " uses" " 
the" " `" "printf" "`" " function" " to" " print" " the" " string" " 
containing" " " "5" " back" "ticks" ":
+ "#+begin_src cpp
+#include <cstdio>
+
+int main() {
+    printf(\"`````\n\");
+    return 0;
+}
+#+end_src")
+
+(gptel-org--test-stream-conversion
+ ;; markdown
+ '("" "Here" " is" " a" " simple" " C" "++" " program" " that" " uses" " the" 
" `" "printf" "`" " function" " to" " print" " the" " string" " containing" " " 
"5" " back" "ticks" ":
 
 " "```" "cpp" "
 " "#include" " <" "iostream" ">
@@ -50,17 +72,8 @@
 " "``" "`
 
 " "In" " this" " code" " snippet" "," " `" "printf" "(\"" "``" "``" "`" "\n" 
"\");" "`" " is" " used" " to" " print" " the" " string" " \"" "``" "```" "\"" 
" followed" " by" " a" " newline" " character" ".")
-        ("" "In" " the" " definition" " of" " the" " `" "struct" " json" 
"_parser" "`," " the" " line" " `" "L" "isp" "_Object" " *" "object" 
"_workspace" ";" "`" " declares" " a" " pointer" " named" " `" "object" 
"_workspace" "`" " of" " type" " `" "L" "isp" "_Object" "`.\n\n" "The" " aster" 
"isk" " (*)" " in" " `" "L" "isp" "_Object" " *" "object" "_workspace" ";" "`" 
" is" " the" " pointer" " ind" "irection" " operator" " in" " C" "." " It" " 
indicates" " that" " `" "object" "_workspa [...]
-      (converted-sequences
-       '("#+begin_src cpp
-#include <cstdio>
-
-int main() {
-    printf(\"`````\n\");
-    return 0;
-}
-#+end_src"
-         "Here is a simple C++ program that uses the =printf= function to 
print the string containing 5 backticks:
+ ;; Org
+  "Here is a simple C++ program that uses the =printf= function to print the 
string containing 5 backticks:
 
 #+begin_src cpp
 #include <iostream>
@@ -73,27 +86,31 @@ int main() {
 }
 #+end_src
 
-In this code snippet, =printf(\"`````\n\");= is used to print the string \"=\" 
followed by a newline character."
-         "In the definition of the =struct json_parser=, the line =Lisp_Object 
*object_workspace;= declares a pointer named =object_workspace= of type 
=Lisp_Object=.
+In this code snippet, =printf(\"`````\n\");= is used to print the string \"=\" 
followed by a newline character.")
 
-The asterisk (*) in =Lisp_Object *object_workspace;= is the pointer 
indirection operator in C. It indicates that =object_workspace= is a pointer to 
an object of type =Lisp_Object=. This means that =object_workspace= will store 
the memory address (location) of a =Lisp_Object= variable rather than storing 
the actual =Lisp_Object= value.
 
-Therefore, =object_workspace= will be used to point to or reference locations 
in memory where =Lisp_Object= instances are stored. This allows the =struct 
json_parser= to store and work with =Lisp_Object= instances indirectly through 
pointers.")))
-  (ert-deftest test--gptel--convert-markdown->org ()
-    (cl-loop
-     for input in string-sequences
-     for output in converted-sequences
-     do
-     (should (string= (gptel--convert-markdown->org (apply #'concat input))
-                      output))))
-
-  (ert-deftest test--gptel--stream-convert-markdown->org ()
-    (cl-loop
-     for input in string-sequences
-     for output in converted-sequences
-     for func = (gptel--stream-convert-markdown->org)
-     do
-     (should
-      (string= (apply #'concat (mapcar func input))
-               output)))))
+(gptel-org--test-stream-conversion
+ ;; markdown
+ '("" "In" " the" " definition" " of" " the" " `" "struct" " json" "_parser" 
"`," " the" " line" " `" "L" "isp" "_Object" " *" "object" "_workspace" ";" "`" 
" declares" " a" " pointer" " named" " `" "object" "_workspace" "`" " of" " 
type" " `" "L" "isp" "_Object" "`.\n\n" "The" " aster" "isk" " (*)" " in" " `" 
"L" "isp" "_Object" " *" "object" "_workspace" ";" "`" " is" " the" " pointer" 
" ind" "irection" " operator" " in" " C" "." " It" " indicates" " that" " `" 
"object" "_workspace" "` [...]
+ ;; org
+ "In the definition of the =struct json_parser=, the line =Lisp_Object 
*object_workspace;= declares a pointer named =object_workspace= of type 
=Lisp_Object=.
+
+The asterisk (*) in =Lisp_Object *object_workspace;= is the pointer 
indirection operator in C. It indicates that =object_workspace= is a pointer to 
an object of type =Lisp_Object=. This means that =object_workspace= will store 
the memory address (location) of a =Lisp_Object= variable rather than storing 
the actual =Lisp_Object= value.
 
+Therefore, =object_workspace= will be used to point to or reference locations 
in memory where =Lisp_Object= instances are stored. This allows the =struct 
json_parser= to store and work with =Lisp_Object= instances indirectly through 
pointers.")
+
+(gptel-org--test-stream-conversion
+ ;; markdown
+ '("#" "# Advantages of"
+   " Org-Mode: A Detailed Overview\n\nOrg-mode is a powerful and versatile 
Emacs extension that offers a wide range of features"
+   " for organizing information, planning projects, and writing documents. 
Here's a detailed overview of its advantages:\n\n**Note Taking and Idea 
Management:**\n\n"
+   "* **Unified platform:** Org-mode provides a single platform for capturing 
notes, ideas, and tasks, eliminating the need for multiple tools and ensuring 
everything is in one place.\n* **Flexibility:** Notes can be structured 
hierarchically"
+   " using headlines and subheadings, allowing for easy organization and 
navigation.\n* **Linking and tags:** Link notes together for easy reference and 
connect them with tags for categorized browsing.\n* **Metadata:** Capture 
additional information like author, deadline, and"
+   " priority for better organization and search.\n\n**Project Management and 
Planning:**\n\n* **Task management:** Create to-do lists with deadlines, 
priority levels, and tags for efficient task management.\n* **Gantt charts and 
time estimates:** Visualize project timelines and estimate time commitments for 
better planning and organization.\n* **Calendar integration:** Link tasks to 
your calendar for better integration and time management.\n* **"
+   "Progress tracking:** Track the progress of tasks and projects with 
checkboxes and progress bars.\n\n**Writing and Document Creation:**\n\n* **Rich 
text formatting:** Org-mode supports a variety of text formatting options, 
including bold, italics, headings, and lists, for creating professional-looking 
documents.\n* **Exporting to various formats:** Easily export your notes and 
documents to various formats like PDF, HTML, LaTeX, and"
+   " plain text.\n* **Beamer presentations:** Create Beamer presentations 
directly within Org-mode for academic or professional presentations.\n* 
**Source code blocks:** Include and highlight code blocks for easy reference 
and documentation.\n\n**Additional Features:**\n\n* **Org-Capture:** Quickly 
capture ideas, notes, and tasks from anywhere using keyboard shortcuts.\n* 
**Org-Agenda:** View and manage your tasks and appointments in a calendar-like 
format.\n* **Emacs Lisp scripting:** A [...]
+   " functionality and automate tasks with Emacs Lisp scripting.\n* **Active 
development and community support:** Benefit from the active development and 
supportive community of Org-mode users.\n\n**Overall Advantages:**\n\n* 
**Increased productivity:** Organize your thoughts, tasks, and projects 
efficiently, leading to increased productivity.\n* **Improved focus:** 
Eliminate distractions and stay focused on the task at hand.\n* **Enhanced 
creativity:** Capture ideas quickly and easily,  [...]
+   " of notes, ideas, and projects for future reference.\n* **Personalized 
workflow:** Tailor Org-mode to your specific needs and preferences for a truly 
personalized workflow.\n\n**Limitations:**\n\n* **Learning curve:** While 
powerful, Org-mode has a steeper learning curve compared to simpler note-taking 
apps.\n* **Emacs dependency:** Org-mode requires Emacs, which may not be 
suitable for all users.\n* **Limited mobile support:** Mobile support for 
Org-mode is available but not as feat [...]
+   " the desktop version.\n\n**Who should use Org-mode?**\n\n* Students\n* 
Researchers\n* Writers\n* Project managers\n* Anyone who wants to improve their 
personal organization and productivity\n\n**Conclusion:**\n\nOrg-mode is a 
powerful and versatile tool that can significantly increase your productivity 
and organization. Its flexibility, rich features, and active community make it 
an excellent choice for managing notes, projects, and documents. If you're 
looking for a way to streamlin [...]
+ ;; org
+ "** Advantages of Org-Mode: A Detailed Overview\n\nOrg-mode is a powerful and 
versatile Emacs extension that offers a wide range of features for organizing 
information, planning projects, and writing documents. Here's a detailed 
overview of its advantages:\n\n*Note Taking and Idea Management:*\n\n- *Unified 
platform:* Org-mode provides a single platform for capturing notes, ideas, and 
tasks, eliminating the need for multiple tools and ensuring everything is in 
one place.\n- *Flexibility [...]