add beginning of hashline verification

1 files changed, 102 insertions, 5 deletions

diff --git a/llm-tools-hl.el b/llm-tools-hl.el
index 5a984e8..1ec13ef 100644
--- a/llm-tools-hl.el
+++ b/llm-tools-hl.el
@@ -89,12 +89,9 @@ If INCLUDE-CONTENT is non-nil, append `|LINE' after each hash."
   "Return FILE contents in hashline format from lines BEG to END."
   (llm-tools--hl-format-lines (llm-tools--hl-file-lines file beg end) beg t))
 
-;; FIXME might be redundant with =llm-tools--hl-file-line-hash= if I'm only
-;; verifying the hash of the anchors and telling LLM to re-read the
-;; file when they don't match
 (defun llm-tools--hl-file-read-hashes (file)
-  "Return FILE contents in hashline format without content."
-  (llm-tools--hl-format-lines (llm-tools--hl-file-lines file)))
+  "TODO rewrite docstring to be less redundant. Return FILE contents in hashline format without content as a list of hashline-formatted hashes."
+  (string-lines (llm-tools--hl-format-lines (llm-tools--hl-file-lines file))))
 
 (defun llm-tools--hl-file-line-hash (file n)
   "Return the hash of the Nth line in FILE."
@@ -107,4 +104,104 @@ If INCLUDE-CONTENT is non-nil, append `|LINE' after each hash."
 ;; rather, the llm would not format it like the reads. it's just going to provide text according to what the anchors it gives say
 ;; so I just need to verify if the anchors didn't change
 
+;;;; Diff Parsing
+(defun llm-tools--split-patch-sections (patch)
+  "Return a list of (FILENAME . REST) cons cells from PATCH.
+Each section starts with '@@ PATH' on the first line; the car is PATH
+and the cdr is the remainder of that section."
+  (let ((pos (string-match "^@@" patch)))
+    (when pos
+      (let (sections
+            (start pos))
+        (while (string-match "\n@@" patch (1+ pos))
+          (let ((nl (match-beginning 0)))
+            (push (substring patch start (1+ nl)) sections)
+            (setq start (1+ nl)))
+          (setq pos (match-end 0)))
+        (push (substring patch start) sections)
+        (mapcar (lambda (sec)
+                  (let ((end (string-match "\n" sec)))
+                    (cons (substring sec 3 end)
+                          (substring sec end))))
+                (nreverse sections))))))
+
+(defun llm-tools--hl-get-anchors (patch)
+  "Returns an alist of (FILENAME . ANCHORS) from PATCH.
+ANCHORS is a sorted, deduplicated list of all line anchors referenced by
+the operations in that file's sections.  Each anchor is a string like
+\"5ff\" or a list of two strings from a range like
+(\"3gv\" \"6be\")."
+  (let ((sections (split-patch-sections patch)))
+    (delete-dups
+     (mapcar (lambda (patch)
+               (cons (car patch)
+                     (delete-dups
+                      (sort
+                       (flatten-list
+                        (mapcar (lambda (line)
+                                  (if (length> line 0)
+                                      (let ((op (char-to-string (elt line 0)))
+                                            (anchor (substring line 2)))
+                                        (pcase op
+                                          ((or "+" "<") anchor)
+                                          ((or "-" "=") (split-string anchor "\\.\\."))))))
+                                (string-lines (cdr patch))))))))
+             sections))))
+
+(defun llm-tools--hl-verify (patch)
+  "TODO rewrite docstring to be lses verbose? Verifies the anchors exist in PATCH."
+  (let ((patch-anchors (llm-tools--hl-get-anchors patch)))
+    (mapcar (lambda (section)
+              (let* ((file (car section))
+                     (anchors (delete "EOF" (cdr section)))
+                     (hl (llm-tools--hl-file-read-hashes file)))
+                ;; cl-subsetp uses eql by default which compares by identity and not content
+                (cons file (cl-subsetp anchors hl :test #'equal))))
+            patch-anchors)))
+
+(let* ((patch "\
+# Replace one line (the payload must re-emit the original indentation)
+@@ mod.ts
+= 1vx..1vx
+~const TITLE = \"Mrs\";
+
+# Replace a full multiline statement (widen to a self-contained boundary)
+@@ mod.ts
+= 3gv..6be
+~       return [
+~               \"Mrs\",
+~               name?.trim() || \"guest\",
+~       ].join(\" \");
+
+# Insert AFTER/BEFORE a line
+@@ mod.ts
++ 4ei
+~               \"Dr\",
+< 5ff
+~               \"Dr\",
+
+# Append to file
+@@ mod.ts
++ EOF
+~export const done = true;
+
+# Delete a line
+@@ test.el
+- 5ff..5ff
+
+# Blank a line (replace with LF)
+@@ mod.ts
+= 5bb..5bb
+"))
+  (string-join
+   (delete-dups
+    (remq nil
+          (mapcar (lambda (elem)
+                    (unless (cdr elem)
+                      (format "%s had incorrect anchors. Use the 'read_file' tool to get correct anchors."
+                              (car elem))))
+                  (llm-tools--hl-verify patch))))
+   "\n"))
+
 (provide 'llm-tools-hl)
+