feat(cli): add metadata inspection mode (#1381)

.github/workflows/build.yml

@@ -64,7 +64,7 @@ jobs:
       - name: Setup pnpm
         uses: pnpm/action-setup@v4
         with:
-          version: 9
+          version: 10.15.1
 
       - name: Dependencies
         id: depends
@@ -127,7 +127,7 @@ jobs:
       - name: Setup pnpm
         uses: pnpm/action-setup@v4
         with:
-          version: 9
+          version: 10.15.1
 
       - name: Dependencies
         id: depends
@@ -205,7 +205,7 @@ jobs:
       - name: Setup pnpm
         uses: pnpm/action-setup@v4
         with:
-          version: 9
+          version: 10.15.1
 
       - name: Get commit hash
         id: commit
@@ -264,7 +264,7 @@ jobs:
       - name: Setup pnpm
         uses: pnpm/action-setup@v4
         with:
-          version: 9
+          version: 10.15.1
 
       - name: Dependencies
         id: depends
@@ -345,7 +345,7 @@ jobs:
       - name: Setup pnpm
         uses: pnpm/action-setup@v4
         with:
-          version: 9
+          version: 10.15.1
 
       - name: Install cuda-toolkit
         id: cuda-toolkit
@@ -460,7 +460,7 @@ jobs:
       - name: Setup pnpm
         uses: pnpm/action-setup@v4
         with:
-          version: 9
+          version: 10.15.1
 
       - name: Cache ROCm Installation
         id: cache-rocm
@@ -573,7 +573,7 @@ jobs:
       - name: Setup pnpm
         uses: pnpm/action-setup@v4
         with:
-          version: 9
+          version: 10.15.1
 
       - name: Free disk space
         run: |

examples/cli/CMakeLists.txt

@@ -1,6 +1,9 @@
 set(TARGET sd-cli)
 
-add_executable(${TARGET} main.cpp)
+add_executable(${TARGET}
+    image_metadata.cpp
+    main.cpp
+)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE stable-diffusion ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE stable-diffusion zip ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PUBLIC c_std_11 cxx_std_17)

examples/cli/README.md

@@ -10,13 +10,18 @@ CLI Options:
   --preview-interval <int>    interval in denoising steps between consecutive updates of the image preview file (default is 1, meaning updating at
                               every step)
   --output-begin-idx <int>    starting index for output image sequence, must be non-negative (default 0 if specified %d in output path, 1 otherwise)
+  --image <string>            path to the image to inspect (for metadata mode)
+  --metadata-format <string>  metadata output format, one of [text, json] (default: text)
   --canny                     apply canny preprocessor (edge detection)
   --convert-name              convert tensor name (for convert mode)
   -v, --verbose               print extra info
   --color                     colors the logging tags according to level
   --taesd-preview-only        prevents usage of taesd for decoding the final image. (for use with --preview tae)
   --preview-noisy             enables previewing noisy inputs of the models rather than the denoised outputs
-  -M, --mode                  run mode, one of [img_gen, vid_gen, upscale, convert], default: img_gen
+  --metadata-raw              include raw hex previews for unparsed metadata payloads
+  --metadata-brief            truncate long metadata text values in text output
+  --metadata-all              include structural/container entries such as IHDR, IDAT, and non-metadata JPEG segments
+  -M, --mode                  run mode, one of [img_gen, vid_gen, upscale, convert, metadata], default: img_gen
   --preview                   preview method. must be one of the following [none, proj, tae, vae] (default is none)
   -h, --help                  show this help message and exit
 
@@ -148,3 +153,12 @@ Generation Options:
   --scm-mask                               SCM steps mask for cache-dit: comma-separated 0/1 (e.g., "1,1,1,0,0,1,0,0,1,0") - 1=compute, 0=can cache
   --scm-policy                             SCM policy: 'dynamic' (default) or 'static'
 ```
+
+Metadata mode inspects PNG/JPEG container metadata without loading any model:
+
+```bash
+./bin/sd-cli -M metadata --image ./output.png
+./bin/sd-cli -M metadata --image ./output.jpg --metadata-format json
+./bin/sd-cli -M metadata --image ./output.png --metadata-raw
+./bin/sd-cli -M metadata --image ./output.png --metadata-all
+```

examples/cli/image_metadata.h

@@ -0,0 +1,21 @@
+#pragma once
+
+#include <iosfwd>
+#include <string>
+
+enum class MetadataOutputFormat {
+    TEXT,
+    JSON,
+};
+
+struct MetadataReadOptions {
+    MetadataOutputFormat output_format = MetadataOutputFormat::TEXT;
+    bool include_raw                   = false;
+    bool brief                         = false;
+    bool include_structural            = false;
+};
+
+bool print_image_metadata(const std::string& image_path,
+                          const MetadataReadOptions& options,
+                          std::ostream& out,
+                          std::string& error);

examples/cli/main.cpp

@@ -18,6 +18,7 @@
 #include "common/common.hpp"
 
 #include "avi_writer.h"
+#include "image_metadata.h"
 
 const char* previews_str[] = {
     "none",
@@ -32,6 +33,8 @@ struct SDCliParams {
     SDMode mode             = IMG_GEN;
     std::string output_path = "output.png";
     int output_begin_idx    = -1;
+    std::string image_path;
+    std::string metadata_format = "text";
 
     bool verbose          = false;
     bool canny_preprocess = false;
@@ -44,6 +47,9 @@ struct SDCliParams {
     bool taesd_preview       = false;
     bool preview_noisy       = false;
     bool color               = false;
+    bool metadata_raw        = false;
+    bool metadata_brief      = false;
+    bool metadata_all        = false;
 
     bool normal_exit = false;
 
@@ -55,6 +61,14 @@ struct SDCliParams {
              "--output",
              "path to write result image to. you can use printf-style %d format specifiers for image sequences (default: ./output.png) (eg. output_%03d.png)",
              &output_path},
+            {"",
+             "--image",
+             "path to the image to inspect (for metadata mode)",
+             &image_path},
+            {"",
+             "--metadata-format",
+             "metadata output format, one of [text, json] (default: text)",
+             &metadata_format},
             {"",
              "--preview-path",
              "path to write preview image to (default: ./preview.png)",
@@ -97,6 +111,18 @@ struct SDCliParams {
              "--preview-noisy",
              "enables previewing noisy inputs of the models rather than the denoised outputs",
              true, &preview_noisy},
+            {"",
+             "--metadata-raw",
+             "include raw hex previews for unparsed metadata payloads",
+             true, &metadata_raw},
+            {"",
+             "--metadata-brief",
+             "truncate long metadata text values in text output",
+             true, &metadata_brief},
+            {"",
+             "--metadata-all",
+             "include structural/container entries such as IHDR, IDAT, and non-metadata JPEG segments",
+             true, &metadata_all},
 
         };
 
@@ -149,7 +175,7 @@ struct SDCliParams {
         options.manual_options = {
             {"-M",
              "--mode",
-             "run mode, one of [img_gen, vid_gen, upscale, convert], default: img_gen",
+             "run mode, one of [img_gen, vid_gen, upscale, convert, metadata], default: img_gen",
              on_mode_arg},
             {"",
              "--preview",
@@ -165,7 +191,7 @@ struct SDCliParams {
     };
 
     bool process_and_check() {
-        if (output_path.length() == 0) {
+        if (mode != METADATA && output_path.length() == 0) {
             LOG_ERROR("error: the following arguments are required: output_path");
             return false;
         }
@@ -174,6 +200,16 @@ struct SDCliParams {
             if (output_path == "output.png") {
                 output_path = "output.gguf";
             }
+        } else if (mode == METADATA) {
+            if (image_path.empty()) {
+                LOG_ERROR("error: metadata mode needs an image path (--image)");
+                return false;
+            }
+            if (metadata_format != "text" && metadata_format != "json") {
+                LOG_ERROR("error: invalid metadata format %s, must be one of [text, json]",
+                          metadata_format.c_str());
+                return false;
+            }
         }
         return true;
     }
@@ -183,6 +219,8 @@ struct SDCliParams {
         oss << "SDCliParams {\n"
             << "  mode: " << modes_str[mode] << ",\n"
             << "  output_path: \"" << output_path << "\",\n"
+            << "  image_path: \"" << image_path << "\",\n"
+            << "  metadata_format: \"" << metadata_format << "\",\n"
             << "  verbose: " << (verbose ? "true" : "false") << ",\n"
             << "  color: " << (color ? "true" : "false") << ",\n"
             << "  canny_preprocess: " << (canny_preprocess ? "true" : "false") << ",\n"
@@ -192,7 +230,10 @@ struct SDCliParams {
             << "  preview_path: \"" << preview_path << "\",\n"
             << "  preview_fps: " << preview_fps << ",\n"
             << "  taesd_preview: " << (taesd_preview ? "true" : "false") << ",\n"
-            << "  preview_noisy: " << (preview_noisy ? "true" : "false") << "\n"
+            << "  preview_noisy: " << (preview_noisy ? "true" : "false") << ",\n"
+            << "  metadata_raw: " << (metadata_raw ? "true" : "false") << ",\n"
+            << "  metadata_brief: " << (metadata_brief ? "true" : "false") << ",\n"
+            << "  metadata_all: " << (metadata_all ? "true" : "false") << "\n"
             << "}";
         return oss.str();
     }
@@ -217,9 +258,13 @@ void parse_args(int argc, const char** argv, SDCliParams& cli_params, SDContextP
         exit(cli_params.normal_exit ? 0 : 1);
     }
 
-    if (!cli_params.process_and_check() ||
+    bool valid = cli_params.process_and_check();
-        !ctx_params.process_and_check(cli_params.mode) ||
+    if (valid && cli_params.mode != METADATA) {
-        !gen_params.process_and_check(cli_params.mode, ctx_params.lora_model_dir)) {
+        valid = ctx_params.process_and_check(cli_params.mode) &&
+                gen_params.process_and_check(cli_params.mode, ctx_params.lora_model_dir);
+    }
+
+    if (!valid) {
         print_usage(argc, argv, options_vec);
         exit(1);
     }
@@ -430,6 +475,27 @@ int main(int argc, const char* argv[]) {
     SDGenerationParams gen_params;
 
     parse_args(argc, argv, cli_params, ctx_params, gen_params);
+    sd_set_log_callback(sd_log_cb, (void*)&cli_params);
+    log_verbose = cli_params.verbose;
+    log_color   = cli_params.color;
+
+    if (cli_params.mode == METADATA) {
+        MetadataReadOptions options;
+        options.output_format      = cli_params.metadata_format == "json"
+                                         ? MetadataOutputFormat::JSON
+                                         : MetadataOutputFormat::TEXT;
+        options.include_raw        = cli_params.metadata_raw;
+        options.brief              = cli_params.metadata_brief;
+        options.include_structural = cli_params.metadata_all;
+
+        std::string error;
+        if (!print_image_metadata(cli_params.image_path, options, std::cout, error)) {
+            LOG_ERROR("%s", error.c_str());
+            return 1;
+        }
+        return 0;
+    }
+
     if (gen_params.video_frames > 4) {
         size_t last_dot_pos   = cli_params.preview_path.find_last_of(".");
         std::string base_path = cli_params.preview_path;
@@ -447,9 +513,6 @@ int main(int argc, const char* argv[]) {
     if (cli_params.preview_method == PREVIEW_PROJ)
         cli_params.preview_fps /= 4;
 
-    sd_set_log_callback(sd_log_cb, (void*)&cli_params);
-    log_verbose = cli_params.verbose;
-    log_color   = cli_params.color;
     sd_set_preview_callback(step_callback,
                             cli_params.preview_method,
                             cli_params.preview_interval,

examples/common/common.hpp

@@ -39,14 +39,16 @@ const char* modes_str[] = {
     "vid_gen",
     "convert",
     "upscale",
+    "metadata",
 };
-#define SD_ALL_MODES_STR "img_gen, vid_gen, convert, upscale"
+#define SD_ALL_MODES_STR "img_gen, vid_gen, convert, upscale, metadata"
 
 enum SDMode {
     IMG_GEN,
     VID_GEN,
     CONVERT,
     UPSCALE,
+    METADATA,
     MODE_COUNT
 };
 
@@ -777,7 +779,7 @@ struct SDContextParams {
     }
 
     bool process_and_check(SDMode mode) {
-        if (mode != UPSCALE && model_path.length() == 0 && diffusion_model_path.length() == 0) {
+        if (mode != UPSCALE && mode != METADATA && model_path.length() == 0 && diffusion_model_path.length() == 0) {
             LOG_ERROR("error: the following arguments are required: model_path/diffusion_model\n");
             return false;
         }