version bump

Fixes #20

.gitmodules

@@ -40,3 +40,6 @@
 [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

@@ -45,7 +45,6 @@ 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 +83,8 @@ target_link_directories(
 target_compile_options(sist2
         PRIVATE
         -Ofast
-        # -march=native
+#        -march=native
+        -fPIC
         -fno-stack-protector
         -fomit-frame-pointer
         )
@@ -111,7 +111,7 @@ TARGET_LINK_LIBRARIES(
         ${PROJECT_SOURCE_DIR}/lib/libonion_static.a
 
         pthread
-        curl
+
         m
         bz2
         ${PROJECT_SOURCE_DIR}/lib/libmagic.a
@@ -128,8 +128,14 @@ TARGET_LINK_LIBRARIES(
         ${PROJECT_SOURCE_DIR}/lib/libtesseract.a
         ${PROJECT_SOURCE_DIR}/lib/liblept.a
         ${PROJECT_SOURCE_DIR}/lib/libtiff.a
-        png
+        ${PROJECT_SOURCE_DIR}/lib/libpng16.a
         stdc++
+
+        # curl
+        ${PROJECT_SOURCE_DIR}/lib/libcurl.a
+        ${PROJECT_SOURCE_DIR}/lib/libcrypto.a
+        ${PROJECT_SOURCE_DIR}/lib/libssl.a
+        dl
 )
 
 add_custom_target(

Docker/Dockerfile

@@ -16,4 +16,7 @@ RUN mkdir -p /usr/share/tessdata && \
 
 ADD sist2 /root/sist2
 
+ENV LANG C.UTF-8
+ENV LC_ALL C.UTF-8
+
 ENTRYPOINT ["/root/sist2"]

Docker/build.sh

@@ -5,6 +5,11 @@ strip sist2
 version=$(./sist2 --version)
 
 echo "Version ${version}"
-docker build . -t simon987/sist2:${version} -t simon987/sist2:latest
+docker build . -t simon987/sist2:${version} -t simon987/sist2:latest \
+  -t docker.pkg.github.com/simon987/sist2/sist2:latest -t docker.pkg.github.com/simon987/sist2/sist2:${version}
 docker push simon987/sist2:${version}
 docker push simon987/sist2:latest
+docker push docker.pkg.github.com/simon987/sist2/sist2:latest
+docker push docker.pkg.github.com/simon987/sist2/sist2:${version}
+
+docker run --rm -it simon987/sist2 -v

README.md

@@ -29,7 +29,7 @@ sist2 (Simple incremental search tool)
 1. Have an [Elasticsearch](https://www.elastic.co/downloads/elasticsearch) instance running
 1. 
     1. Download the [latest sist2 release](https://github.com/simon987/sist2/releases) *
-    1. *(or)* Download an [development snapshot](https://files.simon987.net/artifacts/Sist2/Build/) *(Not recommended!)*
+    1. *(or)* Download a [development snapshot](https://files.simon987.net/artifacts/Sist2/Build/) *(Not recommended!)*
     1. *(or)* `docker pull simon987/sist2:latest`
    
 
@@ -39,6 +39,9 @@ sist2 (Simple incremental search tool)
 
 ## Example usage
 
+
+![demo](demo.gif)
+
 See help page `sist2 --help` for more details.
 
 **Scan a directory**

ci/build.sh

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

schema/mappings.json

@@ -7,7 +7,7 @@
     },
     "suggest-path": {
       "type": "completion",
-      "analyzer": "keyword"
+      "analyzer": "case_insensitive_kw_analyzer"
     },
     "mime": {
       "type": "keyword"

schema/settings.json

@@ -21,6 +21,12 @@
           "lowercase"
         ]
       },
+      "case_insensitive_kw_analyzer": {
+        "tokenizer": "keyword",
+        "filter": [
+          "lowercase"
+        ]
+      },
       "my_nGram": {
         "tokenizer": "my_nGram_tokenizer",
         "filter": [

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 .
@@ -82,7 +83,7 @@ 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
@@ -92,8 +93,36 @@ mv leptonica/src/.libs/liblept.a .
 # tiff
 cd libtiff
 ./autogen.sh
-./configure --enable-static --disable-shared --disable-lzw --disable-jpeg --disable-webp \
+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 .
+mv libtiff/libtiff/.libs/libtiff.a .
+
+# png
+cd libpng
+CFLAGS="-fPIC" ./configure --enable-static --disable-shared
+make -j $THREADS
+cd ..
+mv libpng/.libs/libpng16.a .
+
+# openssl...
+git clone --depth 1 -b OpenSSL_1_1_0-stable https://github.com/openssl/openssl
+cd openssl
+./config --prefix=$(pwd)/../ssl
+make depend
+make -j $THREADS
+make install
+cd ..
+mv ./openssl/libcrypto.a ./openssl/libssl.a .
+
+# curl
+wget -nc https://curl.haxx.se/download/curl-7.68.0.tar.gz
+tar -xzf curl-7.68.0.tar.gz
+cd curl-7.68.0
+./configure --disable-ldap --disable-ldaps --without-librtmp --disable-rtsp --disable-crypto-auth \
+  --disable-smtp --without-libidn2 --without-nghttp2 --without-brotli --enable-static --disable-shared \
+  --without-libpsl --with-ssl=$(pwd)/../ssl
+make -j $THREADS
+cd ..
+mv curl-7.68.0/lib/.libs/libcurl.a .

src/cli.c

@@ -15,6 +15,13 @@
 #define DEFAULT_BIND_ADDR "localhost"
 #define DEFAULT_PORT "4090"
 
+const char* TESS_DATAPATHS[] = {
+        "/usr/share/tessdata/",
+        "/usr/share/tesseract-ocr/tessdata/",
+        "./",
+        NULL
+};
+
 
 scan_args_t *scan_args_create() {
     scan_args_t *args = calloc(sizeof(scan_args_t), 1);
@@ -136,13 +143,23 @@ int scan_args_validate(scan_args_t *args, int argc, const char **argv) {
 
     if (args->tesseract_lang != NULL) {
         TessBaseAPI *api = TessBaseAPICreate();
-        ret = TessBaseAPIInit3(api, TESS_DATAPATH, args->tesseract_lang);
+
+        char filename[128];
+        sprintf(filename, "%s.traineddata", args->tesseract_lang);
+        const char * path = find_file_in_paths(TESS_DATAPATHS, filename);
+        if (path == NULL) {
+            LOG_FATAL("cli.c", "Could not find tesseract language file!");
+        }
+
+        ret = TessBaseAPIInit3(api, path, args->tesseract_lang);
         if (ret != 0) {
             fprintf(stderr, "Could not initialize tesseract with lang '%s'\n", args->tesseract_lang);
             return 1;
         }
         TessBaseAPIEnd(api);
         TessBaseAPIDelete(api);
+
+        args->tesseract_path = path;
     }
 
     LOG_DEBUGF("cli.c", "arg quality=%f", args->quality)
@@ -156,7 +173,8 @@ int scan_args_validate(scan_args_t *args, int argc, const char **argv) {
     LOG_DEBUGF("cli.c", "arg depth=%d", args->depth)
     LOG_DEBUGF("cli.c", "arg path=%s", args->path)
     LOG_DEBUGF("cli.c", "arg archive=%s", args->archive)
-    LOG_DEBUGF("cli.c", "arg ocr=%s", args->tesseract_lang)
+    LOG_DEBUGF("cli.c", "arg tesseract_lang=%s", args->tesseract_lang)
+    LOG_DEBUGF("cli.c", "arg tesseract_path=%s", args->tesseract_path)
 
     return 0;
 }

src/cli.h

@@ -17,6 +17,7 @@ typedef struct scan_args {
     char *archive;
     archive_mode_t archive_mode;
     char *tesseract_lang;
+    const char *tesseract_path;
 } scan_args_t;
 
 scan_args_t *scan_args_create();

src/ctx.h

@@ -28,6 +28,7 @@ struct {
 
     pthread_mutex_t mupdf_mu;
     char * tesseract_lang;
+    char * tesseract_path;
 } ScanCtx;
 
 struct {

src/index/elastic.c

@@ -136,7 +136,7 @@ void elastic_flush() {
         LOG_FATALF("elastic.c", "Could not connect to %s, make sure that elasticsearch is running!\n", IndexCtx.es_url)
     }
 
-    LOG_INFOF("elastic.c", "Indexed %3d documents (%zukB) <%d>", 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) {
@@ -274,3 +274,28 @@ cJSON *elastic_get_document(const char *uuid_str) {
     free_response(r);
     return json;
 }
+
+char *elastic_get_status() {
+    char url[4096];
+    snprintf(url, 4096,
+            "%s/_cluster/state/metadata/sist2?filter_path=metadata.indices.*.state", WebCtx.es_url);
+
+    response_t *r = web_get(url);
+    cJSON *json = NULL;
+    char *status = malloc(128 * sizeof(char));
+    status[0] = '\0';
+
+    if (r->status_code == 200) {
+        json = cJSON_Parse(r->body);
+        const cJSON *metadata = cJSON_GetObjectItem(json, "metadata");
+        if (metadata != NULL) {
+            const cJSON *indices = cJSON_GetObjectItem(metadata, "indices");
+            const cJSON *sist2 = cJSON_GetObjectItem(indices, "sist2");
+            const cJSON *state = cJSON_GetObjectItem(sist2, "state");
+            strcpy(status, state->valuestring);
+        }
+    }
+    free_response(r);
+    cJSON_Delete(json);
+    return status;
+}

src/index/elastic.h

@@ -30,4 +30,6 @@ void elastic_init(int force_reset);
 
 cJSON *elastic_get_document(const char *uuid_str);
 
+char *elastic_get_status();
+
 #endif

src/io/serialize.c

@@ -118,6 +118,22 @@ char *get_meta_key_text(enum metakey meta_key) {
             return "font_name";
         case MetaParent:
             return "parent";
+        case MetaExifMake:
+            return "exif_make";
+        case MetaExifSoftware:
+            return "exif_software";
+        case MetaExifExposureTime:
+            return "exif_exposure_time";
+        case MetaExifFNumber:
+            return "exif_fnumber";
+        case MetaExifFocalLength:
+            return "exif_focal_length";
+        case MetaExifUserComment:
+            return "exif_user_comment";
+        case MetaExifIsoSpeedRatings:
+            return "exif_iso_speed_ratings";
+        case MetaExifModel:
+            return "exif_model";
         default:
             return NULL;
     }
@@ -255,6 +271,14 @@ void read_index_bin(const char *path, const char *index_id, index_func func) {
                 case MetaGenre:
                 case MetaFontName:
                 case MetaParent:
+                case MetaExifMake:
+                case MetaExifSoftware:
+                case MetaExifExposureTime:
+                case MetaExifFNumber:
+                case MetaExifFocalLength:
+                case MetaExifUserComment:
+                case MetaExifIsoSpeedRatings:
+                case MetaExifModel:
                 case MetaTitle: {
                     buf.cur = 0;
                     while ((c = getc(file)) != 0) {

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 = "1.2.1";
+static const char *const Version = "1.2.5";
 static const char *const usage[] = {
         "sist2 scan [OPTION]... PATH",
         "sist2 index [OPTION]... INDEX",
@@ -51,6 +51,7 @@ void sist2_scan(scan_args_t *args) {
     strncpy(ScanCtx.index.desc.root, args->path, sizeof(ScanCtx.index.desc.root));
     ScanCtx.index.desc.root_len = (short) strlen(ScanCtx.index.desc.root);
     ScanCtx.tesseract_lang = args->tesseract_lang;
+    ScanCtx.tesseract_path = args->tesseract_path;
 
     init_dir(ScanCtx.index.path);
 

src/parsing/doc.c

@@ -55,7 +55,14 @@ __always_inline
 void read_part(opcContainer *c, dyn_buffer_t *buf, opcPart part, document_t *doc) {
 
     mceTextReader_t reader;
-    int ret = opcXmlReaderOpen(c, &reader, part, NULL, "UTF-8", 0);
+    int options;
+    if (LogCtx.very_verbose) {
+        options = XML_PARSE_NONET;
+    } else {
+        options = XML_PARSE_NOWARNING | XML_PARSE_NOERROR | XML_PARSE_NONET;
+    }
+
+    int ret = opcXmlReaderOpen(c, &reader, part, NULL, "UTF-8", options);
 
     if (ret != OPC_ERROR_NONE) {
         LOG_ERRORF(doc->filepath, "(doc.c) opcXmlReaderOpen() returned error code %d", ret);

src/parsing/media.c

@@ -193,6 +193,22 @@ append_video_meta(AVFormatContext *pFormatCtx, AVFrame *frame, document_t *doc,
                 APPEND_TAG_META(doc, tag, MetaArtist)
             } else if (strcmp(tag->key, "ImageDescription") == 0) {
                 APPEND_TAG_META(doc, tag, MetaContent)
+            } else if (strcmp(tag->key, "Make") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifMake)
+            } else if (strcmp(tag->key, "Model") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifModel)
+            } else if (strcmp(tag->key, "Software") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifSoftware)
+            } else if (strcmp(tag->key, "FNumber") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifFNumber)
+            } else if (strcmp(tag->key, "FocalLength") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifFocalLength)
+            } else if (strcmp(tag->key, "UserComment") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifUserComment)
+            } else if (strcmp(tag->key, "ISOSpeedRatings") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifIsoSpeedRatings)
+            } else if (strcmp(tag->key, "ExposureTime") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifExposureTime)
             }
         }
     }

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;
 
@@ -140,7 +141,7 @@ void fill_image(fz_context *ctx, UNUSED(fz_device *dev),
 
         if (pix->h > MIN_OCR_SIZE && img->h > MIN_OCR_SIZE && img->xres != 0) {
             TessBaseAPI *api = TessBaseAPICreate();
-            TessBaseAPIInit3(api, TESS_DATAPATH, ScanCtx.tesseract_lang);
+            TessBaseAPIInit3(api, ScanCtx.tesseract_path, ScanCtx.tesseract_lang);
 
             TessBaseAPISetImage(api, pix->samples, pix->w, pix->h, pix->n, pix->stride);
             TessBaseAPISetSourceResolution(api, pix->xres);
@@ -156,8 +157,8 @@ void fill_image(fz_context *ctx, UNUSED(fz_device *dev),
 
             TessBaseAPIEnd(api);
             TessBaseAPIDelete(api);
-            fz_drop_pixmap(ctx, pix);
         }
+        fz_drop_pixmap(ctx, pix);
     }
 }
 

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

@@ -3,7 +3,6 @@
 
 #define UUID_STR_LEN 37
 #define UNUSED(x) __attribute__((__unused__))  x
-#define TESS_DATAPATH "/usr/share/tessdata/"
 
 #include <glib-2.0/glib.h>
 #include <unistd.h>
@@ -34,6 +33,8 @@
 #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

@@ -2,9 +2,9 @@
 #define SIST2_TYPES_H
 
 
-#define META_INT_MASK 0xF0
-#define META_STR_MASK 0xE0
-#define META_LONG_MASK 0xD0
+#define META_INT_MASK 0x80
+#define META_STR_MASK 0x40
+#define META_LONG_MASK 0x20
 #define IS_META_INT(key) (key & META_INT_MASK) == META_INT_MASK
 #define IS_META_LONG(key) (key & META_LONG_MASK) == META_LONG_MASK
 #define IS_META_STR(meta) (meta->key & META_STR_MASK) == META_STR_MASK
@@ -31,6 +31,15 @@ enum metakey {
     MetaTitle = 12 | META_STR_MASK,
     MetaFontName = 13 | META_STR_MASK,
     MetaParent = 14 | META_STR_MASK,
+    MetaExifMake = 15 | META_STR_MASK,
+    MetaExifSoftware = 16 | META_STR_MASK,
+    MetaExifExposureTime = 17 | META_STR_MASK,
+    MetaExifFNumber = 18 | META_STR_MASK,
+    MetaExifFocalLength = 19 | META_STR_MASK,
+    MetaExifUserComment = 20 | META_STR_MASK,
+    MetaExifModel = 21 | META_STR_MASK,
+    MetaExifIsoSpeedRatings = 22 | META_STR_MASK,
+    //Note to self: this will break after 31 entries
 };
 
 #define INDEX_TYPE_BIN "binary"

src/util.c

@@ -1,4 +1,5 @@
 #include "util.h"
+#include "src/ctx.h"
 
 dyn_buffer_t dyn_buffer_create() {
     dyn_buffer_t buf;
@@ -90,7 +91,11 @@ text_buffer_t text_buffer_create(int max_size) {
 }
 
 void text_buffer_terminate_string(text_buffer_t *buf) {
-    dyn_buffer_write_char(&buf->dyn_buffer, '\0');
+    if (*(buf->dyn_buffer.buf + buf->dyn_buffer.cur - 1) == ' ') {
+        *(buf->dyn_buffer.buf + buf->dyn_buffer.cur - 1) = '\0';
+    } else {
+        dyn_buffer_write_char(&buf->dyn_buffer, '\0');
+    }
 }
 
 __always_inline
@@ -171,8 +176,8 @@ int text_buffer_append_string0(text_buffer_t *buf, char *str) {
 
 int text_buffer_append_char(text_buffer_t *buf, int c) {
 
-    if (SHOULD_IGNORE_CHAR(c)) {
-        if (!buf->last_char_was_whitespace) {
+    if (SHOULD_IGNORE_CHAR(c) || c == ' ') {
+        if (!buf->last_char_was_whitespace && buf->dyn_buffer.cur != 0) {
             dyn_buffer_write_char(&buf->dyn_buffer, ' ');
             buf->last_char_was_whitespace = TRUE;
 
@@ -317,4 +322,29 @@ GHashTable *incremental_get_table() {
     return file_table;
 }
 
+const char *find_file_in_paths(const char *paths[], const char *filename) {
+
+    for (int i = 0; paths[i] != NULL; i++) {
+
+        char *apath = abspath(paths[i]);
+        if (apath == NULL) {
+            continue;
+        }
+
+        char path[PATH_MAX];
+        snprintf(path, sizeof(path), "%s%s", apath, filename);
+
+        LOG_DEBUGF("util.c", "Looking for '%s' in folder '%s'", filename, apath)
+        free(apath);
+
+        struct stat info;
+        int ret = stat(path, &info);
+        if (ret != -1) {
+            return paths[i];
+        }
+    }
+
+    return NULL;
+}
+
 

src/util.h

@@ -74,5 +74,6 @@ int incremental_get(GHashTable *table, unsigned long inode_no);
 
 int incremental_mark_file_for_copy(GHashTable *table, unsigned long inode_no);
 
+const char *find_file_in_paths(const char **paths, const char *filename);
 
 #endif

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;
@@ -398,6 +431,23 @@ int file(void *p, onion_request *req, onion_response *res) {
     return ret;
 }
 
+int status(UNUSED(void *p), UNUSED(onion_request *req), onion_response *res) {
+    set_default_headers(res);
+
+    onion_response_set_header(res, "Content-Type", "application/x-empty");
+
+    char *status = elastic_get_status();
+    if (strcmp(status, "open") == 0) {
+        onion_response_set_code(res, 204);
+    } else {
+        onion_response_set_code(res, 500);
+    }
+
+    free(status);
+
+    return OCS_PROCESSED;
+}
+
 void serve(const char *hostname, const char *port) {
     onion *o = onion_new(O_POOL);
     onion_set_timeout(o, 3500);
@@ -417,6 +467,7 @@ void serve(const char *hostname, const char *port) {
 
     onion_url_add(urls, "es", search);
     onion_url_add(urls, "scroll", scroll);
+    onion_url_add(urls, "status", status);
     onion_url_add(
             urls,
             "^t/([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})/"
@@ -424,6 +475,7 @@ 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">v1.2.1</span>
+    <span class="badge badge-pill version">v1.2.5</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>