mtmd : add const in various places (#28307)

* mtmd : mark context as const in more methods

Mark `mtmd_context` as `const` in:
- mtmd_bitmap_init_lazy
- mtmd_tokenize
- mtmd_tokenize_from_parts
- mtmd_helper_support_video
- mtmd_helper_bitmap_init_from_file
- mtmd_helper_bitmap_init_from_buf
- mtmd_helper_video_init
- mtmd_helper_video_init_from_buf
- mtmd_helper_model_can_chat

The tokenization functions in particular are useful to have marked
`const`, as that allows more easily telling the compiler that we can
safely tokenize from multiple threads (`mtmd_tokenize` is already
documented as thread-safe, this just reifies that in the signature).

* mtmd : mark tokenization input pointer as const

Mark the `bitmaps` and `parts` pointers in `mtmd_tokenize` and
`mtmd_tokenize_from_parts` as `const`. This allows more easily calling
these with immutable arrays / vectors.

* mtmd : mark llama_context as const in mtmd_helper_model_can_chat

tools/mtmd/mtmd-helper.cpp

@@ -369,11 +369,11 @@ static bool is_webp_file(const unsigned char * buf, size_t len) {
 }
 
 #ifdef MTMD_VIDEO
-static mtmd_bitmap * decode_webp_with_ffmpeg(mtmd_context * mctx, const unsigned char * buf, size_t len, bool placeholder,
+static mtmd_bitmap * decode_webp_with_ffmpeg(const mtmd_context * mctx, const unsigned char * buf, size_t len, bool placeholder,
                                              const mtmd_helper_video_init_params & params);
 #endif
 
-mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_buf(mtmd_context * ctx, const unsigned char * buf, size_t len, bool placeholder,
+mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_buf(const mtmd_context * ctx, const unsigned char * buf, size_t len, bool placeholder,
                                                             mtmd_helper_init_opt opt) {
     // calculate the hash if needed
     std::string id;
@@ -459,7 +459,7 @@ mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_buf(mtmd_context * ctx,
     return {nullptr, nullptr};
 }
 
-mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_file(mtmd_context * ctx, const char * fname, bool placeholder,
+mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_file(const mtmd_context * ctx, const char * fname, bool placeholder,
                                                              mtmd_helper_init_opt opt) {
 #ifdef _WIN32
     int wlen = MultiByteToWideChar(CP_UTF8, 0, fname, -1, NULL, 0);
@@ -504,7 +504,7 @@ mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_file(mtmd_context * ctx,
     return mtmd_helper_bitmap_init_from_buf(ctx, buf.data(), buf.size(), placeholder, opt);
 }
 
-bool mtmd_helper_support_video(mtmd_context * ctx) {
+bool mtmd_helper_support_video(const mtmd_context * ctx) {
 #ifdef MTMD_VIDEO
     return mtmd_support_vision(ctx);
 #else
@@ -520,7 +520,7 @@ bool mtmd_helper_support_video(mtmd_context * ctx) {
 #ifdef MTMD_VIDEO
 
 struct mtmd_helper_video {
-    mtmd_context * mctx;
+    const mtmd_context * mctx;
     std::string path;
     std::vector<uint8_t> input_buf; // non-empty when initialized from buffer
     std::string ffmpeg_bin;
@@ -886,7 +886,7 @@ static std::string video_resolve_bin(const char * bin_dir, const char * name) {
 }
 
 #ifdef MTMD_VIDEO
-static mtmd_bitmap * decode_webp_with_ffmpeg(mtmd_context * mctx, const unsigned char * buf, size_t len, bool placeholder,
+static mtmd_bitmap * decode_webp_with_ffmpeg(const mtmd_context * mctx, const unsigned char * buf, size_t len, bool placeholder,
                                              const mtmd_helper_video_init_params & params) {
     mtmd_helper_video vctx;
     vctx.mctx        = mctx;
@@ -913,7 +913,7 @@ static mtmd_bitmap * decode_webp_with_ffmpeg(mtmd_context * mctx, const unsigned
 #endif
 
 mtmd_helper_video * mtmd_helper_video_init(
-        mtmd_context * mctx,
+        const mtmd_context * mctx,
         const char * path,
         mtmd_helper_video_init_params params) {
 #ifdef MTMD_VIDEO
@@ -948,7 +948,7 @@ mtmd_helper_video * mtmd_helper_video_init(
 }
 
 mtmd_helper_video * mtmd_helper_video_init_from_buf(
-        mtmd_context * mctx,
+        const mtmd_context * mctx,
         const unsigned char * buf, size_t len,
         mtmd_helper_video_init_params params) {
 #ifdef MTMD_VIDEO
@@ -1016,7 +1016,7 @@ int32_t mtmd_helper_video_read_next(mtmd_helper_video * ctx,
 #endif
 }
 
-bool mtmd_helper_model_can_chat(llama_context * lctx, mtmd_context * mctx) {
+bool mtmd_helper_model_can_chat(const llama_context * lctx, const mtmd_context * mctx) {
     if (!mctx) {
         return true;
     }

tools/mtmd/mtmd-helper.h

@@ -46,7 +46,7 @@ MTMD_API struct mtmd_helper_init_opt mtmd_helper_init_opt_default(void);
 MTMD_API void mtmd_helper_log_set(ggml_log_callback log_callback, void * user_data);
 
 // Returns true if this build includes video support (MTMD_VIDEO was ON at compile time).
-MTMD_API bool mtmd_helper_support_video(mtmd_context * ctx);
+MTMD_API bool mtmd_helper_support_video(const mtmd_context * ctx);
 
 struct mtmd_helper_bitmap_wrapper {
     mtmd_bitmap * bitmap;
@@ -58,7 +58,7 @@ struct mtmd_helper_bitmap_wrapper {
 // returns nullptr on failure
 // this function is thread-safe
 MTMD_API struct mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_file(
-                    mtmd_context * ctx,
+                    const mtmd_context * ctx,
                     const char * fname,
                     bool placeholder,
                     struct mtmd_helper_init_opt opt);
@@ -75,7 +75,7 @@ MTMD_API struct mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_file(
 // returns nullptr on failure
 // this function is thread-safe
 MTMD_API struct mtmd_helper_bitmap_wrapper mtmd_helper_bitmap_init_from_buf(
-                    mtmd_context * ctx,
+                    const mtmd_context * ctx,
                     const unsigned char * buf, size_t len,
                     bool placeholder,
                     struct mtmd_helper_init_opt opt);
@@ -153,7 +153,7 @@ struct mtmd_helper_video_info {
 
 // returns NULL on failure (ffprobe not found, file unreadable, etc.)
 MTMD_API mtmd_helper_video * mtmd_helper_video_init(
-                    struct mtmd_context * mctx,
+                    const struct mtmd_context * mctx,
                     const char * path,
                     struct mtmd_helper_video_init_params params);
 
@@ -162,7 +162,7 @@ MTMD_API mtmd_helper_video * mtmd_helper_video_init(
 // Note: pipe input is not seekable, so seeking will use output-side seeking
 // (ffmpeg decodes and discards frames up to the target position).
 MTMD_API mtmd_helper_video * mtmd_helper_video_init_from_buf(
-                    struct mtmd_context * mctx,
+                    const struct mtmd_context * mctx,
                     const unsigned char * buf, size_t len,
                     struct mtmd_helper_video_init_params params);
 MTMD_API void mtmd_helper_video_free(mtmd_helper_video * ctx);
@@ -177,7 +177,7 @@ MTMD_API int32_t mtmd_helper_video_read_next(mtmd_helper_video * ctx,
             char ** out_text);
 
 // return true if model can be used for chat
-MTMD_API bool mtmd_helper_model_can_chat(struct llama_context * lctx, struct mtmd_context * mctx);
+MTMD_API bool mtmd_helper_model_can_chat(const struct llama_context * lctx, const struct mtmd_context * mctx);
 
 //
 // Audio generation helpers

tools/mtmd/mtmd.cpp

@@ -1117,7 +1117,7 @@ std::vector<std::vector<const mtmd_bitmap *>> mtmd_group_mergeable_bitmaps(std::
 }
 
 struct mtmd_tokenizer {
-    mtmd_context * ctx;
+    const mtmd_context * ctx;
 
     std::string input_text; // note: can contain null bytes; do not use c_str()
     bool add_special;
@@ -1140,9 +1140,9 @@ struct mtmd_tokenizer {
         }
     }
 
-    mtmd_tokenizer(mtmd_context * ctx,
+    mtmd_tokenizer(const mtmd_context * ctx,
             const mtmd_input_text * text,
-            const mtmd_bitmap ** bmps,
+            const mtmd_bitmap * const * bmps,
             size_t n_bitmaps) : ctx(ctx) {
         add_special   = text->add_special;
         parse_special = text->parse_special;
@@ -1177,8 +1177,8 @@ struct mtmd_tokenizer {
         expand_lazy_bitmaps();
     }
 
-    mtmd_tokenizer(mtmd_context * ctx,
-            const mtmd_input_part ** input_parts,
+    mtmd_tokenizer(const mtmd_context * ctx,
+            const mtmd_input_part * const * input_parts,
             size_t n_parts,
             bool add_special) : ctx(ctx) {
         this->add_special = add_special;
@@ -1733,10 +1733,10 @@ struct mtmd_tokenizer {
     }
 };
 
-int32_t mtmd_tokenize(mtmd_context * ctx,
+int32_t mtmd_tokenize(const mtmd_context * ctx,
             mtmd_input_chunks * output,
             const mtmd_input_text * text,
-            const mtmd_bitmap ** bitmaps,
+            const mtmd_bitmap * const * bitmaps,
             size_t n_bitmaps) {
     try {
         mtmd_tokenizer tokenizer(ctx, text, bitmaps, n_bitmaps);
@@ -1747,9 +1747,9 @@ int32_t mtmd_tokenize(mtmd_context * ctx,
     }
 }
 
-int32_t mtmd_tokenize_from_parts(mtmd_context * ctx,
+int32_t mtmd_tokenize_from_parts(const mtmd_context * ctx,
             mtmd_input_chunks * output,
-            const mtmd_input_part ** parts,
+            const mtmd_input_part * const * parts,
             size_t n_parts,
             bool add_special) {
     for (size_t i = 0; i < n_parts; i++) {
@@ -2271,7 +2271,7 @@ void mtmd_bitmap_set_mergeable(mtmd_bitmap * bitmap, bool mergeable) {
     bitmap->mergeable = mergeable;
 }
 
-mtmd_bitmap * mtmd_bitmap_init_lazy(mtmd_context * ctx,
+mtmd_bitmap * mtmd_bitmap_init_lazy(const mtmd_context * ctx,
                                     const char * id,
                                     void * user_data,
                                     mtmd_bitmap_lazy_callback callback) {

tools/mtmd/mtmd.h

@@ -211,7 +211,7 @@ typedef int(* mtmd_bitmap_lazy_callback)(
     mtmd_bitmap ** out_bitmap,
     char ** out_text);
 
-MTMD_API mtmd_bitmap * mtmd_bitmap_init_lazy(mtmd_context * ctx,
+MTMD_API mtmd_bitmap * mtmd_bitmap_init_lazy(const mtmd_context * ctx,
                                              const char * id, // usually set to file hash
                                              void * user_data,
                                              mtmd_bitmap_lazy_callback callback);
@@ -299,10 +299,10 @@ MTMD_API struct mtmd_decoder_pos mtmd_image_tokens_get_decoder_pos(const mtmd_im
 //   0 on success
 //   1 on number of bitmaps not matching the number of markers
 //   2 on media preprocessing error
-MTMD_API int32_t mtmd_tokenize(mtmd_context * ctx,
+MTMD_API int32_t mtmd_tokenize(const mtmd_context * ctx,
                                mtmd_input_chunks * output,
                                const mtmd_input_text * text,
-                               const mtmd_bitmap ** bitmaps,
+                               const mtmd_bitmap * const * bitmaps,
                                size_t n_bitmaps);
 
 // same as mtmd_tokenize(), but takes an array of mtmd_input_part
@@ -311,9 +311,9 @@ MTMD_API int32_t mtmd_tokenize(mtmd_context * ctx,
 // - when you want to control parse_special for each text part
 // note: per-part add_special will be ignored
 // return 1 if a part has both text and bitmap set (or neither)
-MTMD_API int32_t mtmd_tokenize_from_parts(mtmd_context * ctx,
+MTMD_API int32_t mtmd_tokenize_from_parts(const mtmd_context * ctx,
                                           mtmd_input_chunks * output,
-                                          const mtmd_input_part ** parts,
+                                          const mtmd_input_part * const * parts,
                                           size_t n_parts,
                                           bool add_special);