Bundle libcurl attempt 2

.gitmodules

@@ -37,3 +37,9 @@
 [submodule "lib/leptonica"]
 	path = lib/leptonica
 	url = https://github.com/danbloomberg/leptonica
+[submodule "lib/libtiff"]
+	path = lib/libtiff
+	url = https://gitlab.com/libtiff/libtiff
+[submodule "lib/libpng"]
+	path = lib/libpng
+	url = https://github.com/glennrp/libpng

CMakeLists.txt

@@ -1,6 +1,8 @@
 cmake_minimum_required(VERSION 3.7)
 set(CMAKE_C_STANDARD 11)
 
+option(CURL_STATIC "Link to curl statically" on)
+
 project(sist2 C)
 list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_LIST_DIR}/CMakeModules")
 
@@ -45,7 +47,6 @@ add_executable(
 find_package(PkgConfig REQUIRED)
 set(ENV{PKG_CONFIG_PATH} "$ENV{PKG_CONFIG_PATH}:/usr/local/lib/pkgconfig/")
 
-#find_package(OpenSSL REQUIRED)
 find_package(Freetype REQUIRED)
 
 pkg_check_modules(GLIB REQUIRED glib-2.0)
@@ -84,7 +85,8 @@ target_link_directories(
 target_compile_options(sist2
         PRIVATE
         -Ofast
-        # -march=native
+        -march=native
+        -fPIC
         -fno-stack-protector
         -fomit-frame-pointer
         )
@@ -111,7 +113,11 @@ TARGET_LINK_LIBRARIES(
         ${PROJECT_SOURCE_DIR}/lib/libonion_static.a
 
         pthread
-        curl
+
+        # curl
+        ${PROJECT_SOURCE_DIR}/lib/libcurl.a
+        ssl crypto
+
         m
         bz2
         ${PROJECT_SOURCE_DIR}/lib/libmagic.a
@@ -127,8 +133,8 @@ TARGET_LINK_LIBRARIES(
 
         ${PROJECT_SOURCE_DIR}/lib/libtesseract.a
         ${PROJECT_SOURCE_DIR}/lib/liblept.a
-        png
-        tiff
+        ${PROJECT_SOURCE_DIR}/lib/libtiff.a
+        ${PROJECT_SOURCE_DIR}/lib/libpng16.a
         stdc++
 )
 

README.md

@@ -143,7 +143,7 @@ binaries.
     apt install git cmake pkg-config libglib2.0-dev \
         libssl-dev uuid-dev python3 libmagic-dev libfreetype6-dev \
         libcurl-dev libbz2-dev yasm libharfbuzz-dev ragel \
-        libarchive-dev libtiff5 libpng16-16
+        libarchive-dev libtiff5 libpng16-16 libpango1.0-dev
    ```
 
 2. Build

ci/build.sh

@@ -5,4 +5,3 @@
 cmake .
 make
 strip sist2
-strip sist2_scan

schema/mappings.json

@@ -13,19 +13,24 @@
       "type": "keyword"
     },
     "videoc": {
-      "type": "keyword"
+      "type": "keyword",
+      "index": false
     },
     "audioc": {
-      "type": "keyword"
+      "type": "keyword",
+      "index": false
     },
     "duration": {
-      "type": "float"
+      "type": "float",
+      "index": false
     },
     "width": {
-      "type": "integer"
+      "type": "integer",
+      "index": false
     },
     "height": {
-      "type": "integer"
+      "type": "integer",
+      "index": false
     },
     "mtime": {
       "type": "integer"
@@ -70,6 +75,23 @@
       "analyzer": "my_nGram",
       "type": "text"
     },
+    "_keyword.*":  {
+      "type": "keyword"
+    },
+    "_text.*":  {
+      "analyzer": "content_analyzer",
+      "type": "text",
+      "fields": {
+        "nGram": {
+          "type": "text",
+          "analyzer": "my_nGram"
+        }
+      }
+    },
+    "_url": {
+      "type": "keyword",
+      "index": false
+    },
     "content": {
       "analyzer": "content_analyzer",
       "type": "text",

scripts/get_static_libs.sh

@@ -5,7 +5,7 @@ THREADS=$(nproc)
 cd lib
 
 cd mupdf
-make USE_SYSTEM_HARFBUZZ=yes USE_SYSTEM_OPENJPEG=yes HAVE_X11=no HAVE_GLUT=no -j $THREADS
+CFLAGS=-fPIC make USE_SYSTEM_HARFBUZZ=yes USE_SYSTEM_OPENJPEG=yes HAVE_X11=no HAVE_GLUT=no -j $THREADS
 cd ..
 
 mv mupdf/build/release/libmupdf.a .
@@ -13,8 +13,7 @@ mv mupdf/build/release/libmupdf-third.a .
 
 # openjp2
 cd openjpeg
-#cmake . -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_FLAGS="-O3 -march=native -DNDEBUG"
-cmake . -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_FLAGS="-O3"
+cmake . -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_FLAGS="-O3 -march=native -DNDEBUG -fPIC"
 make -j $THREADS
 cd ..
 mv openjpeg/bin/libopenjp2.a .
@@ -22,7 +21,7 @@ mv openjpeg/bin/libopenjp2.a .
 # harfbuzz
 cd harfbuzz
 ./autogen.sh
-./configure --disable-shared --enable-static
+CFLAGS=-fPIC ./configure --disable-shared --enable-static
 make -j $THREADS
 cd ..
 mv harfbuzz/src/.libs/libharfbuzz.a .
@@ -33,7 +32,8 @@ cd ffmpeg
  --disable-ffprobe --disable-doc\
  --disable-manpages --disable-postproc --disable-avfilter \
  --disable-alsa --disable-lzma --disable-xlib --disable-debug\
- --disable-vdpau --disable-vaapi --disable-sdl2 --disable-network
+ --disable-vdpau --disable-vaapi --disable-sdl2 --disable-network\
+ --extra-cflags=-fPIC
 make -j $THREADS
 cd ..
 
@@ -74,7 +74,8 @@ mv libmagic/src/.libs/libmagic.a .
 cd tesseract
 mkdir build
 cd build
-cmake -DSTATIC=on -DBUILD_TRAINING_TOOLS=off ..
+cmake -DSTATIC=on -DBUILD_TRAINING_TOOLS=off -DBUILD_TESTS=off -DCMAKE_BUILD_TYPE=Release \
+  -DCMAKE_CXX_FLAGS="-fPIC"  ..
 make -j $THREADS
 cd ../..
 mv tesseract/build/libtesseract.a .
@@ -82,9 +83,35 @@ mv tesseract/build/libtesseract.a .
 # leptonica
 cd leptonica
 ./autogen.sh
-./configure --without-zlib --without-jpeg --without-giflib \
+CFLAGS="-fPIC" ./configure --without-zlib --without-jpeg --without-giflib \
   --without-giflib --without-libwebp --without-libwebpmux --without-libopenjpeg \
   --enable-static --disable-shared
 make -j $THREADS
 cd ..
 mv leptonica/src/.libs/liblept.a .
+
+# tiff
+cd libtiff
+./autogen.sh
+CFLAGS="-fPIC" CXXFLAGS="-fPIC" CXX_FLAGS="-fPIC" ./configure --enable-static --disable-shared --disable-lzw --disable-jpeg --disable-webp \
+  --disable-lzma --disable-zstd --disable-jbig
+make -j $THREADS
+cd ..
+mv libtiff/libtiff/.libs/libtiff.a .
+
+# png
+cd libpng
+CFLAGS="-fPIC" ./configure --enable-static --disable-shared
+make -j $THREADS
+cd ..
+mv libpng/.libs/libpng16.a .
+
+# curl
+wget -nc https://curl.haxx.se/download/curl-7.68.0.tar.gz
+tar -xzf curl-7.68.0.tar.gz
+cd curl-7.68.0
+./configure --disable-ldap --disable-ldaps --without-librtmp --disable-rtsp --disable-crypto-auth \
+  --disable-smtp --enable-static --disable-shared
+make -j $THREADS
+cd ..
+mv curl-7.68.0/lib/.libs/libcurl.a .

scripts/get_static_libs_freebsd.sh

@@ -1,58 +0,0 @@
-#!/usr/bin/env bash
-cd lib
-
-# mupdf
-cd mupdf
-HAVE_X11=no HAVE_GLUT=no gmake -j 4
-cd ..
-
-mv mupdf/build/release/libmupdf.a .
-mv mupdf/build/release/libmupdf-third.a .
-
-# openjp2
-cd openjpeg
-#cmake . -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_FLAGS="-O3 -march=native -DNDEBUG"
-cmake . -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_FLAGS="-O3"
-gmake -j 4
-cd ..
-mv openjpeg/bin/libopenjp2.a .
-
-# harfbuzz
-cd harfbuzz
-./autogen.sh
-./configure --disable-shared --enable-static
-gmake -j 4
-cd ..
-mv harfbuzz/src/.libs/libharfbuzz.a .
-
-# ffmpeg
-cd ffmpeg
-./configure --disable-shared --enable-static --disable-ffmpeg --disable-ffplay \
- --disable-ffprobe --disable-doc\
- --disable-manpages --disable-postproc --disable-avfilter \
- --disable-alsa --disable-lzma --disable-xlib --disable-debug\
- --disable-vdpau --disable-vaapi --disable-sdl2 --disable-network
-gmake -j 4
-cd ..
-
-mv ffmpeg/libavcodec/libavcodec.a .
-mv ffmpeg/libavformat/libavformat.a .
-mv ffmpeg/libavutil/libavutil.a .
-mv ffmpeg/libswresample/libswresample.a .
-mv ffmpeg/libswscale/libswscale.a .
-
-#bzip2
-cd bzip2-1.0.6
-make -j 4
-cd ..
-mv bzip2-1.0.6/libbz2.a .
-
-# magic
-cd libmagic
-./autogen.sh
-./configure --enable-static --disable-shared
-make -j 4
-cd ..
-mv libmagic/src/.libs/libmagic.a .
-
-cd ..

src/cli.c

@@ -163,6 +163,8 @@ int scan_args_validate(scan_args_t *args, int argc, const char **argv) {
 
 int index_args_validate(index_args_t *args, int argc, const char **argv) {
 
+    LogCtx.verbose = 1;
+
     if (argc < 2) {
         fprintf(stderr, "Required positional argument: PATH.\n");
         return 1;
@@ -224,6 +226,8 @@ int index_args_validate(index_args_t *args, int argc, const char **argv) {
 
 int web_args_validate(web_args_t *args, int argc, const char **argv) {
 
+    LogCtx.verbose = 1;
+
     if (argc < 2) {
         fprintf(stderr, "Required positional argument: PATH.\n");
         return 1;

src/index/elastic.c

@@ -27,13 +27,14 @@ void print_json(cJSON *document, const char uuid_str[UUID_STR_LEN]) {
     cJSON_AddStringToObject(line, "_id", uuid_str);
     cJSON_AddStringToObject(line, "_index", "sist2");
     cJSON_AddStringToObject(line, "_type", "_doc");
-    cJSON_AddItemToObject(line, "_source", document);
+    cJSON_AddItemReferenceToObject(line, "_source", document);
 
     char *json = cJSON_PrintUnformatted(line);
 
     printf("%s\n", json);
 
-    cJSON_free(line);
+    cJSON_free(json);
+    cJSON_Delete(line);
 }
 
 void index_json(cJSON *document, const char uuid_str[UUID_STR_LEN]) {
@@ -68,7 +69,7 @@ void execute_update_script(const char *script, const char index_id[UUID_STR_LEN]
     char bulk_url[4096];
     snprintf(bulk_url, 4096, "%s/sist2/_update_by_query?pretty", Indexer->es_url);
     response_t *r = web_post(bulk_url, str, "Content-Type: application/json");
-    printf("Executed user script <%d>\n", r->status_code);
+    LOG_INFOF("elastic.c", "Executed user script <%d>", r->status_code);
     cJSON *resp = cJSON_Parse(r->body);
 
     cJSON_free(str);
@@ -79,7 +80,7 @@ void execute_update_script(const char *script, const char index_id[UUID_STR_LEN]
     if (error != NULL) {
         char *error_str = cJSON_Print(error);
 
-        fprintf(stderr, "User script error: \n%s\n", error_str);
+        LOG_ERRORF("elastic.c", "User script error: \n%s", error_str);
         cJSON_free(error_str);
     }
 
@@ -132,11 +133,10 @@ void elastic_flush() {
     response_t *r = web_post(bulk_url, buf, "Content-Type: application/x-ndjson");
 
     if (r->status_code == 0) {
-        fprintf(stderr, "Could not connect to %s, make sure that elasticsearch is running!\n", IndexCtx.es_url);
-        exit(1);
+        LOG_FATALF("elastic.c", "Could not connect to %s, make sure that elasticsearch is running!\n", IndexCtx.es_url)
     }
 
-    printf("Indexed %3d documents (%zukB) <%d>\n", count, buf_cur / 1024, r->status_code);
+    LOG_INFOF("elastic.c", "Indexed %d documents (%zukB) <%d>", count, buf_cur / 1024, r->status_code);
 
     cJSON *ret_json = cJSON_Parse(r->body);
     if (cJSON_GetObjectItem(ret_json, "errors")->valueint != 0) {
@@ -144,7 +144,7 @@ void elastic_flush() {
         cJSON_ArrayForEach(err, cJSON_GetObjectItem(ret_json, "items")) {
             if (cJSON_GetObjectItem(cJSON_GetObjectItem(err, "index"), "status")->valueint != 201) {
                 char* str = cJSON_Print(err);
-                fprintf(stderr, "%s\n", str);
+                LOG_ERRORF("elastic.c", "%s\n", str);
                 cJSON_free(str);
             }
         }
@@ -198,7 +198,7 @@ void destroy_indexer(char * script, char index_id[UUID_STR_LEN]) {
 
     snprintf(url, sizeof(url), "%s/sist2/_refresh", IndexCtx.es_url);
     response_t *r = web_post(url, "", NULL);
-    printf("Refresh index <%d>\n", r->status_code);
+    LOG_INFOF("elastic.c", "Refresh index <%d>", r->status_code);
     free_response(r);
 
     if (script != NULL) {
@@ -207,12 +207,12 @@ void destroy_indexer(char * script, char index_id[UUID_STR_LEN]) {
 
     snprintf(url, sizeof(url), "%s/sist2/_refresh", IndexCtx.es_url);
     r = web_post(url, "", NULL);
-    printf("Refresh index <%d>\n", r->status_code);
+    LOG_INFOF("elastic.c", "Refresh index <%d>", r->status_code);
     free_response(r);
 
     snprintf(url, sizeof(url), "%s/sist2/_forcemerge", IndexCtx.es_url);
     r = web_post(url, "", NULL);
-    printf("Merge index <%d>\n", r->status_code);
+    LOG_INFOF("elastic.c", "Merge index <%d>", r->status_code);
     free_response(r);
 
     if (Indexer != NULL) {
@@ -232,32 +232,32 @@ void elastic_init(int force_reset) {
 
     if (!index_exists || force_reset) {
         r = web_delete(url);
-        printf("Delete index <%d>\n", r->status_code);
+        LOG_INFOF("elastic.c", "Delete index <%d>", r->status_code);
         free_response(r);
 
         snprintf(url, 4096, "%s/sist2", IndexCtx.es_url);
         r = web_put(url, "", NULL);
-        printf("Create index <%d>\n", r->status_code);
+        LOG_INFOF("elastic.c", "Create index <%d>", r->status_code);
         free_response(r);
 
         snprintf(url, 4096, "%s/sist2/_close", IndexCtx.es_url);
         r = web_post(url, "", NULL);
-        printf("Close index <%d>\n", r->status_code);
+        LOG_INFOF("elastic.c", "Close index <%d>", r->status_code);
         free_response(r);
 
         snprintf(url, 4096, "%s/sist2/_settings", IndexCtx.es_url);
         r = web_put(url, settings_json, "Content-Type: application/json");
-        printf("Update settings <%d>\n", r->status_code);
+        LOG_INFOF("elastic.c", "Update settings <%d>", r->status_code);
         free_response(r);
 
         snprintf(url, 4096, "%s/sist2/_mappings/_doc?include_type_name=true", IndexCtx.es_url);
         r = web_put(url, mappings_json, "Content-Type: application/json");
-        printf("Update mappings <%d>\n", r->status_code);
+        LOG_INFOF("elastic.c", "Update mappings <%d>", r->status_code);
         free_response(r);
 
         snprintf(url, 4096, "%s/sist2/_open", IndexCtx.es_url);
         r = web_post(url, "", NULL);
-        printf("Open index <%d>\n", r->status_code);
+        LOG_INFOF("elastic.c", "Open index <%d>", r->status_code);
         free_response(r);
     }
 }

src/io/serialize.c

@@ -34,6 +34,7 @@ void write_index_descriptor(char *path, index_descriptor_t *desc) {
     cJSON_AddStringToObject(json, "version", desc->version);
     cJSON_AddStringToObject(json, "root", desc->root);
     cJSON_AddStringToObject(json, "name", desc->name);
+    cJSON_AddStringToObject(json, "type", desc->type);
     cJSON_AddStringToObject(json, "rewrite_url", desc->rewrite_url);
     cJSON_AddNumberToObject(json, "timestamp", (double) desc->timestamp);
 
@@ -56,8 +57,7 @@ index_descriptor_t read_index_descriptor(char *path) {
     int fd = open(path, O_RDONLY);
 
     if (fd == -1) {
-        fprintf(stderr, "Invalid/corrupt index (Could not find descriptor)\n");
-        exit(1);
+        LOG_FATAL("serialize.c", "Invalid/corrupt index (Could not find descriptor)\n")
     }
 
     char *buf = malloc(info.st_size + 1);
@@ -75,6 +75,11 @@ index_descriptor_t read_index_descriptor(char *path) {
     descriptor.root_len = (short) strlen(descriptor.root);
     strcpy(descriptor.version, cJSON_GetObjectItem(json, "version")->valuestring);
     strcpy(descriptor.uuid, cJSON_GetObjectItem(json, "uuid")->valuestring);
+    if (cJSON_GetObjectItem(json, "type") == NULL) {
+        strcpy(descriptor.type, INDEX_TYPE_BIN);
+    } else {
+        strcpy(descriptor.type, cJSON_GetObjectItem(json, "type")->valuestring);
+    }
 
     cJSON_Delete(json);
     free(buf);
@@ -172,8 +177,8 @@ void thread_cleanup() {
     close(index_fd);
 }
 
-void read_index(const char *path, const char index_id[UUID_STR_LEN], index_func func) {
 
+void read_index_bin(const char *path, const char *index_id, index_func func) {
     line_t line;
     dyn_buffer_t buf = dyn_buffer_create();
 
@@ -229,7 +234,7 @@ void read_index(const char *path, const char index_id[UUID_STR_LEN], index_func
                 case MetaMediaBitrate: {
                     long value;
                     fread(&value, sizeof(long), 1, file);
-                    cJSON_AddNumberToObject(document, get_meta_key_text(key), value);
+                    cJSON_AddNumberToObject(document, get_meta_key_text(key), (double) value);
                     break;
                 }
                 case MetaMediaAudioCodec:
@@ -262,7 +267,7 @@ void read_index(const char *path, const char index_id[UUID_STR_LEN], index_func
                     break;
                 }
                 default:
-                    LOG_FATALF("serialize.c", "Invalid meta key (corrupt index): %x", key)
+                LOG_FATALF("serialize.c", "Invalid meta key (corrupt index): %x", key)
             }
 
             key = getc(file);
@@ -275,6 +280,89 @@ void read_index(const char *path, const char index_id[UUID_STR_LEN], index_func
     fclose(file);
 }
 
+const char *json_type_copy_fields[] = {
+        "mime", "name", "path", "extension", "index", "size", "mtime", "parent",
+
+        // Meta
+        "title", "content", "width", "height", "duration", "audioc", "videoc",
+        "bitrate", "artist", "album", "album_artist", "genre", "title", "font_name",
+
+        // Special
+        "tag", "_url"
+};
+
+const char *json_type_array_fields[] = {
+        "_keyword", "_text"
+};
+
+void read_index_json(const char *path, UNUSED(const char *index_id), index_func func) {
+
+    FILE *file = fopen(path, "r");
+    while (1) {
+        char *line = NULL;
+        size_t len;
+        size_t read = getline(&line, &len, file);
+        if (read == -1) {
+            if (line) {
+                free(line);
+            }
+            break;
+        }
+
+        cJSON *input = cJSON_Parse(line);
+        if (input == NULL) {
+            LOG_FATALF("serialize.c", "Could not parse JSON line: \n%s", line)
+        }
+        if (line) {
+            free(line);
+        }
+
+        cJSON *document = cJSON_CreateObject();
+        const char *uuid_str = cJSON_GetObjectItem(input, "_id")->valuestring;
+
+        for (int i = 0; i < (sizeof(json_type_copy_fields) / sizeof(json_type_copy_fields[0])); i++) {
+            cJSON *value = cJSON_GetObjectItem(input, json_type_copy_fields[i]);
+            if (value != NULL) {
+                cJSON_AddItemReferenceToObject(document, json_type_copy_fields[i], value);
+            }
+        }
+
+        for (int i = 0; i < (sizeof(json_type_array_fields) / sizeof(json_type_array_fields[0])); i++) {
+            cJSON *arr = cJSON_GetObjectItem(input, json_type_array_fields[i]);
+            if (arr != NULL) {
+                cJSON *obj;
+                cJSON_ArrayForEach(obj, arr) {
+                    char key[1024];
+                    cJSON *k = cJSON_GetObjectItem(obj, "k");
+                    cJSON *v = cJSON_GetObjectItem(obj, "v");
+                    if (k == NULL || v == NULL || !cJSON_IsString(k) || !cJSON_IsString(v)) {
+                        char *str = cJSON_Print(obj);
+                        LOG_FATALF("serialize.c", "Invalid %s member: must contain .k and .v string fields: \n%s",
+                                   json_type_array_fields[i], str)
+                    }
+                    snprintf(key, sizeof(key), "%s.%s", json_type_array_fields[i], k->valuestring);
+                    cJSON_AddStringToObject(document, key, v->valuestring);
+                }
+            }
+        }
+
+        func(document, uuid_str);
+        cJSON_Delete(document);
+        cJSON_Delete(input);
+
+    }
+    fclose(file);
+}
+
+void read_index(const char *path, const char index_id[UUID_STR_LEN], const char *type, index_func func) {
+
+    if (strcmp(type, INDEX_TYPE_BIN) == 0) {
+        read_index_bin(path, index_id, func);
+    } else if (strcmp(type, INDEX_TYPE_JSON) == 0) {
+        read_index_json(path, index_id, func);
+    }
+}
+
 void incremental_read(GHashTable *table, const char *filepath) {
     FILE *file = fopen(filepath, "rb");
     line_t line;

src/io/serialize.h

@@ -11,7 +11,7 @@ void incremental_copy(store_t *store, store_t *dst_store, const char *filepath,
 
 void write_document(document_t *doc);
 
-void read_index(const char *path, const char[UUID_STR_LEN], index_func);
+void read_index(const char *path, const char[UUID_STR_LEN], const char *type, index_func);
 
 void incremental_read(GHashTable *table, const char *filepath);
 

src/log.c

@@ -56,7 +56,7 @@ void sist_logf(char *filepath, int level, char *format, ...) {
         log_len += 1;
     }
 
-    write(STDOUT_FILENO, log_str, log_len);
+    write(STDERR_FILENO, log_str, log_len);
 }
 
 void sist_log(char *filepath, int level, char *str) {

src/main.c

@@ -6,7 +6,7 @@
 #define EPILOG "Made by simon987 <me@simon987.net>. Released under GPL-3.0"
 
 
-static const char *const Version = "1.2.0";
+static const char *const Version = "1.2.3";
 static const char *const usage[] = {
         "sist2 scan [OPTION]... PATH",
         "sist2 index [OPTION]... INDEX",
@@ -29,6 +29,7 @@ void init_dir(const char *dirpath) {
     uuid_unparse(uuid, ScanCtx.index.desc.uuid);
     time(&ScanCtx.index.desc.timestamp);
     strcpy(ScanCtx.index.desc.version, Version);
+    strcpy(ScanCtx.index.desc.type, INDEX_TYPE_BIN);
 
     write_index_descriptor(path, &ScanCtx.index.desc);
 }
@@ -130,8 +131,12 @@ void sist2_index(index_args_t *args) {
     snprintf(descriptor_path, PATH_MAX, "%s/descriptor.json", args->index_path);
 
     index_descriptor_t desc = read_index_descriptor(descriptor_path);
-    if (strcmp(desc.version, Version) != 0) {
-        fprintf(stderr, "Version mismatch! Index is v%s but executable is v%s\n", desc.version, Version);
+
+    LOG_DEBUGF("main.c", "descriptor version %s (%s)", desc.version, desc.type)
+
+    if (strcmp(desc.version, Version) != 0 && strcmp(desc.version, INDEX_VERSION_EXTERNAL) != 0) {
+        fprintf(stderr, "Version mismatch! Index is %s but executable is %s/%s\n",
+                desc.version, Version, INDEX_VERSION_EXTERNAL);
         return;
     }
 
@@ -153,7 +158,7 @@ void sist2_index(index_args_t *args) {
         if (strncmp(de->d_name, "_index_", sizeof("_index_") - 1) == 0) {
             char file_path[PATH_MAX];
             snprintf(file_path, PATH_MAX, "%s/%s", args->index_path, de->d_name);
-            read_index(file_path, desc.uuid, f);
+            read_index(file_path, desc.uuid, desc.type, f);
         }
     }
     closedir(dir);

src/parsing/pdf.c

@@ -82,7 +82,8 @@ fz_page *render_cover(fz_context *ctx, document_t *doc, fz_document *fzdoc) {
     fz_drop_pixmap(ctx, pixmap);
 
     if (err != 0) {
-        LOG_WARNINGF(doc->filepath, "fz_new_buffer_from_pixmap_as_png() returned error code [%d] %s", err, ctx->error.message)
+        LOG_WARNINGF(doc->filepath, "fz_new_buffer_from_pixmap_as_png() returned error code [%d] %s", err,
+                     ctx->error.message)
         fz_drop_page(ctx, cover);
         return NULL;
     }
@@ -129,8 +130,8 @@ int read_stext_block(fz_stext_block *block, text_buffer_t *tex) {
 
 
 void fill_image(fz_context *ctx, UNUSED(fz_device *dev),
-        fz_image *img, UNUSED(fz_matrix ctm), UNUSED(float alpha),
-        UNUSED(fz_color_params color_params)) {
+                fz_image *img, UNUSED(fz_matrix ctm), UNUSED(float alpha),
+                UNUSED(fz_color_params color_params)) {
 
     int l2factor = 0;
 

src/parsing/pdf.h

@@ -2,7 +2,6 @@
 #define SIST2_PDF_H
 
 #include "src/sist.h"
-#include <tesseract/capi.h>
 
 
 void parse_pdf(void *buf, size_t buf_len, document_t *doc);

src/sist.h

@@ -34,6 +34,8 @@
 #include <archive_entry.h>
 #include <opc/opc.h>
 #include <libxml/xmlstring.h>
+#define BOOL int
+#include <tesseract/capi.h>
 
 #include <onion/onion.h>
 #include <onion/handler.h>

src/types.h

@@ -33,6 +33,10 @@ enum metakey {
     MetaParent = 14 | META_STR_MASK,
 };
 
+#define INDEX_TYPE_BIN "binary"
+#define INDEX_TYPE_JSON "json"
+#define INDEX_VERSION_EXTERNAL "_external_v1"
+
 typedef struct index_descriptor {
     char uuid[UUID_STR_LEN];
     char version[6];
@@ -41,6 +45,7 @@ typedef struct index_descriptor {
     char rewrite_url[8196];
     short root_len;
     char name[1024];
+    char type[64];
 } index_descriptor_t;
 
 typedef struct index_t {

src/web/serve.c

@@ -110,7 +110,7 @@ int thumbnail(void *p, onion_request *req, onion_response *res) {
     int written = onion_response_write(res, data, data_len);
     onion_response_flush(res);
     if (written != data_len || data_len == 0) {
-        printf("Couldn't write thumb\n");
+        LOG_DEBUG("serve.c", "Couldn't write thumbnail");
     }
     free(data);
 
@@ -214,7 +214,7 @@ int chunked_response_file(const char *filename, const char *mime,
     return OCS_PROCESSED;
 }
 
-int search(void *p, onion_request *req, onion_response *res) {
+int search(UNUSED(void *p), onion_request *req, onion_response *res) {
 
     int flags = onion_request_get_flags(req);
     if ((flags & OR_METHODS) != OR_POST) {
@@ -254,7 +254,7 @@ int search(void *p, onion_request *req, onion_response *res) {
     return OCS_PROCESSED;
 }
 
-int scroll(void *p, onion_request *req, onion_response *res) {
+int scroll(UNUSED(void *p), onion_request *req, onion_response *res) {
 
     int flags = onion_request_get_flags(req);
     if ((flags & OR_METHODS) != OR_GET) {
@@ -327,7 +327,7 @@ int serve_file_from_disk(cJSON *json, index_t *idx, onion_request *req, onion_re
     return chunked_response_file(full_path, mime, 1, req, res);
 }
 
-int index_info(void *p, onion_request *req, onion_response *res) {
+int index_info(UNUSED(void *p), onion_request *req, onion_response *res) {
     cJSON *json = cJSON_CreateObject();
     cJSON *arr = cJSON_AddArrayToObject(json, "indices");
 
@@ -353,14 +353,47 @@ int index_info(void *p, onion_request *req, onion_response *res) {
     return OCS_PROCESSED;
 }
 
-int file(void *p, onion_request *req, onion_response *res) {
+
+int document_info(UNUSED(void *p), onion_request *req, onion_response *res) {
 
     const char *arg_uuid = onion_request_get_query(req, "1");
     if (arg_uuid == NULL) {
         return OCS_PROCESSED;
     }
 
-    char *next = arg_uuid;
+    cJSON *doc = elastic_get_document(arg_uuid);
+    cJSON *source = cJSON_GetObjectItem(doc, "_source");
+
+    cJSON *index_id = cJSON_GetObjectItem(source, "index");
+    if (index_id == NULL) {
+        cJSON_Delete(doc);
+        return OCS_NOT_PROCESSED;
+    }
+
+    index_t *idx = get_index_by_id(index_id->valuestring);
+    if (idx == NULL) {
+        cJSON_Delete(doc);
+        return OCS_NOT_PROCESSED;
+    }
+
+    onion_response_set_header(res, "Content-Type", "application/json");
+
+    char *json_str = cJSON_PrintUnformatted(source);
+    onion_response_write0(res, json_str);
+    free(json_str);
+    cJSON_Delete(doc);
+
+    return OCS_PROCESSED;
+}
+
+int file(UNUSED(void *p), onion_request *req, onion_response *res) {
+
+    const char *arg_uuid = onion_request_get_query(req, "1");
+    if (arg_uuid == NULL) {
+        return OCS_PROCESSED;
+    }
+
+    const char *next = arg_uuid;
     cJSON *doc = NULL;
     cJSON *index_id = NULL;
     cJSON *source = NULL;
@@ -424,6 +457,7 @@ void serve(const char *hostname, const char *port) {
             thumbnail
     );
     onion_url_add(urls, "^f/([a-fA-F0-9]{8}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{12})$", file);
+    onion_url_add(urls, "^d/([a-fA-F0-9]{8}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{12})$", document_info);
     onion_url_add(urls, "i", index_info);
 
 

web/css/dark.css

@@ -2,6 +2,26 @@
     outline: 0;
 }
 
+.info-icon {
+    width: 1rem;
+    margin-right: 0.2rem;
+    cursor: pointer;
+    color: #757575;
+    line-height: 1rem;
+    height: 1.1rem;
+}
+
+.info-icon:hover {
+    color: inherit;
+}
+
+.modal-title {
+    max-width: calc(100% - 2rem);
+    overflow: hidden;
+    text-overflow: ellipsis;
+}
+
+
 .path-row {
     display: -ms-flexbox;
     display: flex;
@@ -32,7 +52,7 @@ body {
     margin-top: 1em;
 }
 
-.card {
+.card, .modal-content {
     margin-top: 1em;
     background: #212121;
     color: #e0e0e0;
@@ -40,6 +60,27 @@ body {
     border: none;
 }
 
+.table {
+    color: #e0e0e0;
+}
+
+.table td, .table th {
+    border: none;
+}
+
+.table thead th {
+    border-bottom: 1px solid #646464;
+}
+
+.modal-header .close {
+    color: #e0e0e0;
+    text-shadow: none;
+}
+
+.modal-header {
+    border-bottom: 1px solid #646464;
+}
+
 .sub-document {
     background: #37474F !important;
 }
@@ -134,6 +175,8 @@ body {
 
 .file-title {
     width: 100%;
+    line-height: 1rem;
+    height: 1.1rem;
     font-size: 10pt;
     white-space: nowrap;
     text-overflow: ellipsis;

web/css/light.css

@@ -2,6 +2,25 @@
     outline: 0;
 }
 
+.info-icon {
+    width: 1rem;
+    margin-right: 0.2rem;
+    cursor: pointer;
+    color: #757575;
+    line-height: 1rem;
+    height: 1rem;
+}
+
+.info-icon:hover {
+    color: inherit;
+}
+
+.modal-title {
+    max-width: calc(100% - 2rem);
+    overflow: hidden;
+    text-overflow: ellipsis;
+}
+
 .path-row {
     display: -ms-flexbox;
     display: flex;
@@ -100,6 +119,8 @@ body {
 
 .file-title {
     width: 100%;
+    line-height: 1rem;
+    height: 1.1rem;
     font-size: 10pt;
     white-space: nowrap;
     text-overflow: ellipsis;

web/js/dom.js

@@ -153,11 +153,44 @@ function getTags(hit, mimeCategory) {
     return tags
 }
 
-/**
- *
- * @param hit
- * @returns {Element}
- */
+function infoButtonCb(hit) {
+    return () => {
+        getDocumentInfo(hit["_id"]).then(doc => {
+            $("#modal-title").text(doc["name"] + (doc["extension"] ? "." + doc["extension"] : ""));
+
+            const tbody = $("<tbody>");
+            $("#modal-body").empty()
+                .append($("<table class='table table-sm'>")
+                    .append($("<thead>")
+                        .append($("<tr>")
+                            .append($("<th>").text("Field"))
+                            .append($("<th>").text("Value"))
+                        )
+                    )
+                    .append(tbody)
+                );
+
+            const displayFields = new Set([
+                "mime", "size", "mtime", "path", "title", "width", "height", "duration", "audioc", "videoc",
+                "bitrate", "artist", "album", "album_artist", "genre", "title", "font_name", "tag"
+            ]);
+            Object.keys(doc)
+                .filter(key => key.startsWith("_keyword.") || key.startsWith("_text.") || displayFields.has(key))
+                .forEach(key => {
+                    tbody.append($("<tr>")
+                        .append($("<td>").text(key))
+                        .append($("<td>").text(doc[key]))
+                    );
+                });
+            if (doc.hasOwnProperty("content") && doc["content"]) {
+                $("#modal-body").append($("<div class='content-div'>").text(doc["content"]))
+            }
+
+            $("#modal").modal();
+        });
+    }
+}
+
 function createDocCard(hit) {
     let docCard = document.createElement("div");
     docCard.setAttribute("class", "card");
@@ -172,6 +205,7 @@ function createDocCard(hit) {
     let link = document.createElement("a");
     link.setAttribute("href", "f/" + hit["_id"]);
     link.setAttribute("target", "_blank");
+    link.style.maxWidth = "calc(100% - 1.2rem)";
     link.appendChild(title);
 
     if (hit["_source"].hasOwnProperty("parent")) {
@@ -271,7 +305,15 @@ function createDocCard(hit) {
     sizeTag.setAttribute("class", "text-muted");
     tagContainer.appendChild(sizeTag);
 
-    docCardBody.appendChild(link);
+    const titleWrapper = document.createElement("div");
+    titleWrapper.style.display = "flex";
+
+    const infoButton = makeInfoButton(hit);
+
+    titleWrapper.appendChild(infoButton);
+    titleWrapper.appendChild(link);
+
+    docCardBody.appendChild(titleWrapper);
     docCard.appendChild(docCardBody);
 
     docCardBody.appendChild(tagContainer);
@@ -352,6 +394,14 @@ function makeThumbnail(mimeCategory, hit, imgWrapper, small) {
     return thumbnail;
 }
 
+function makeInfoButton(hit) {
+    const infoButton = document.createElement("span");
+    infoButton.appendChild(document.createTextNode("🛈"));
+    infoButton.setAttribute("class", "info-icon");
+    infoButton.addEventListener("click", infoButtonCb(hit));
+    return infoButton;
+}
+
 function createDocLine(hit) {
 
     const mime = hit["_source"]["mime"];
@@ -372,6 +422,8 @@ function createDocLine(hit) {
         isSubDocument = true;
     }
 
+    const infoButton = makeInfoButton(hit);
+
     const title = makeTitle(hit);
 
     let link = document.createElement("a");
@@ -380,8 +432,13 @@ function createDocLine(hit) {
     link.appendChild(title);
 
     const titleDiv = document.createElement("div");
-    titleDiv.setAttribute("class", "file-title");
-    titleDiv.appendChild(link);
+
+    const titleWrapper = document.createElement("div");
+    titleWrapper.style.display = "flex";
+    titleWrapper.appendChild(infoButton);
+    titleWrapper.appendChild(link);
+
+    titleDiv.appendChild(titleWrapper);
 
     line.appendChild(media);
 

web/js/search.js

@@ -81,6 +81,13 @@ $.jsonPost("i").then(resp => {
     });
 });
 
+function getDocumentInfo(id) {
+    return $.getJSON("d/" + id).fail(e => {
+        console.log(e);
+        showEsError();
+    })
+}
+
 function handleTreeClick(tree) {
     return (event, node, handler) => {
         event.preventTreeDefault();

web/search.html

@@ -11,7 +11,7 @@
 
 <nav class="navbar navbar-expand-lg">
     <a class="navbar-brand" href="/">sist2</a>
-    <span class="badge badge-pill version">v1.2.0</span>
+    <span class="badge badge-pill version">v1.2.3</span>
     <span class="tagline">Lightning-fast file system indexer and search tool </span>
     <a style="margin-left: auto" id="theme" class="btn" title="Toggle theme" href="/">Theme</a>
 </nav>
@@ -65,6 +65,20 @@
         </div>
     </div>
 
+    <div class="modal" id="modal" tabindex="-1" role="dialog" aria-labelledby="modal-title" aria-hidden="true">
+        <div class="modal-dialog modal-lg modal-dialog-centered" role="document">
+            <div class="modal-content">
+                <div class="modal-header">
+                    <h5 class="modal-title" id="modal-title"></h5>
+                    <button type="button" class="close" data-dismiss="modal" aria-label="Close">
+                        <span aria-hidden="true">&times;</span>
+                    </button>
+                </div>
+                <div class="modal-body" id="modal-body"></div>
+            </div>
+        </div>
+    </div>
+
     <div id="searchResults"></div>
 </div>