add exif dateTime, allow some special characters in text meta

Fixes #20

CMakeLists.txt

@@ -1,8 +1,6 @@
 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")
 
@@ -85,7 +83,7 @@ target_link_directories(
 target_compile_options(sist2
         PRIVATE
         -Ofast
-        -march=native
+#        -march=native
         -fPIC
         -fno-stack-protector
         -fomit-frame-pointer
@@ -114,10 +112,6 @@ TARGET_LINK_LIBRARIES(
 
         pthread
 
-        # curl
-        ${PROJECT_SOURCE_DIR}/lib/libcurl.a
-        ssl crypto
-
         m
         bz2
         ${PROJECT_SOURCE_DIR}/lib/libmagic.a
@@ -136,6 +130,12 @@ TARGET_LINK_LIBRARIES(
         ${PROJECT_SOURCE_DIR}/lib/libtiff.a
         ${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**
@@ -92,7 +95,7 @@ File type | Library | Content | Thumbnail | Metadata
 pdf,xps,cbz,fb2,epub | MuPDF | text+ocr | yes, `png` | title |
 `audio/*` | ffmpeg | - | yes, `jpeg` | ID3 tags |
 `video/*` | ffmpeg | - | yes, `jpeg` | title, comment, artist |
-`image/*` | ffmpeg | - | yes, `jpeg` | `EXIF:Artist`, `EXIF:ImageDescription` |
+`image/*` | ffmpeg | - | yes, `jpeg` | [Common EXIF tags](https://github.com/simon987/sist2/blob/efdde2734eca9b14a54f84568863b7ffd59bdba3/src/parsing/media.c#L190) |
 ttf,ttc,cff,woff,fnt,otf | Freetype2 | - | yes, `bmp` | Name & style |
 `text/plain` | *(none)* | yes | no | - |
 tar, zip, rar, 7z, ar ...  | Libarchive | yes\* | - | no |

schema/mappings.json

@@ -1,5 +1,9 @@
 {
   "properties": {
+    "_tie": {
+      "type": "keyword",
+      "doc_values": true
+    },
     "path": {
       "type": "text",
       "analyzer": "path_analyzer",
@@ -7,7 +11,7 @@
     },
     "suggest-path": {
       "type": "completion",
-      "analyzer": "keyword"
+      "analyzer": "case_insensitive_kw_analyzer"
     },
     "mime": {
       "type": "keyword"
@@ -105,6 +109,30 @@
     },
     "tag": {
       "type": "keyword"
+    },
+    "exif_make": {
+      "type": "text"
+    },
+    "exif_model": {
+      "type": "text"
+    },
+    "exif:software": {
+      "type": "text"
+    },
+    "exif_exposure_time": {
+      "type": "keyword"
+    },
+    "exif_fnumber": {
+      "type": "keyword"
+    },
+    "exif_iso_speed_ratings": {
+      "type": "keyword"
+    },
+    "exif_focal_length": {
+      "type": "keyword"
+    },
+    "exif_user_comment": {
+      "type": "text"
     }
   }
 }

schema/pipeline.json

@@ -0,0 +1,10 @@
+{
+  "description": "Copy _id to _tie",
+  "processors": [
+    {
+      "script": {
+        "source": "ctx._tie = ctx._id;"
+      }
+    }
+  ]
+}

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

@@ -75,7 +75,7 @@ cd tesseract
 mkdir build
 cd build
 cmake -DSTATIC=on -DBUILD_TRAINING_TOOLS=off -DBUILD_TESTS=off -DCMAKE_BUILD_TYPE=Release \
-  -DCMAKE_CXX_FLAGS="-fPIC"  ..
+  -DCMAKE_CXX_FLAGS="-fPIC" -DAUTO_OPTIMIZE=off ..
 make -j $THREADS
 cd ../..
 mv tesseract/build/libtesseract.a .
@@ -106,12 +106,23 @@ 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 $THREADS
+make install
+cd ..
+mv ./openssl/libcrypto.a ./openssl/libssl.a .
+
 # curl
 wget -nc https://curl.haxx.se/download/curl-7.68.0.tar.gz
 tar -xzf curl-7.68.0.tar.gz
 cd curl-7.68.0
 ./configure --disable-ldap --disable-ldaps --without-librtmp --disable-rtsp --disable-crypto-auth \
-  --disable-smtp --enable-static --disable-shared
+  --disable-smtp --without-libidn2 --without-nghttp2 --without-brotli --enable-static --disable-shared \
+  --without-libpsl --with-ssl=$(pwd)/../ssl
 make -j $THREADS
 cd ..
 mv curl-7.68.0/lib/.libs/libcurl.a .

scripts/index_static.py

@@ -1,6 +1,9 @@
+import json
+
 files = [
     "schema/mappings.json",
     "schema/settings.json",
+    "schema/pipeline.json",
 ]
 
 
@@ -9,6 +12,6 @@ def clean(filepath):
 
 
 for file in files:
-    with open(file, "rb") as f:
-        data = f.read()
+    with open(file, "r") as f:
+        data = json.dumps(json.load(f), separators=(",", ":")).encode()
     print("char %s[%d] = {%s};" % (clean(file), len(data), ",".join(str(int(b)) for b in data)))

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

@@ -129,7 +129,7 @@ void elastic_flush() {
     Indexer->queued = 0;
 
     char bulk_url[4096];
-    snprintf(bulk_url, 4096, "%s/sist2/_bulk", Indexer->es_url);
+    snprintf(bulk_url, 4096, "%s/sist2/_bulk?pipeline=tie", Indexer->es_url);
     response_t *r = web_post(bulk_url, buf, "Content-Type: application/x-ndjson");
 
     if (r->status_code == 0) {
@@ -245,6 +245,11 @@ void elastic_init(int force_reset) {
         LOG_INFOF("elastic.c", "Close index <%d>", r->status_code);
         free_response(r);
 
+        snprintf(url, 4096, "%s/_ingest/pipeline/tie", IndexCtx.es_url);
+        r = web_put(url, pipeline_json, "Content-Type: application/json");
+        LOG_INFOF("elastic.c", "Create pipeline <%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");
         LOG_INFOF("elastic.c", "Update settings <%d>", r->status_code);
@@ -274,3 +279,28 @@ 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 +118,24 @@ 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";
+        case MetaExifDateTime:
+            return "exif_datetime";
         default:
             return NULL;
     }
@@ -255,6 +273,15 @@ 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 MetaExifDateTime:
+                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 = "1.2.3";
+static const char *const Version = "1.2.9";
 static const char *const usage[] = {
         "sist2 scan [OPTION]... PATH",
         "sist2 index [OPTION]... INDEX",
@@ -49,8 +49,10 @@ void sist2_scan(scan_args_t *args) {
     strncpy(ScanCtx.index.path, args->output, sizeof(ScanCtx.index.path));
     strncpy(ScanCtx.index.desc.name, args->name, sizeof(ScanCtx.index.desc.name));
     strncpy(ScanCtx.index.desc.root, args->path, sizeof(ScanCtx.index.desc.root));
+    strncpy(ScanCtx.index.desc.rewrite_url, args->rewrite_url, sizeof(ScanCtx.index.desc.rewrite_url));
     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

@@ -1,10 +1,20 @@
 #include "doc.h"
 #include "src/ctx.h"
 
-void dump_text(mceTextReader_t *reader, dyn_buffer_t *buf) {
+int dump_text(mceTextReader_t *reader, dyn_buffer_t *buf) {
 
     mce_skip_attributes(reader);
 
+    xmlErrorPtr err = xmlGetLastError();
+    if (err != NULL) {
+        if (err->level == XML_ERR_FATAL) {
+            LOG_ERRORF("doc.c", "Got fatal XML error while parsing document: %s", err->message)
+            return -1;
+        } else {
+            LOG_ERRORF("doc.c", "Got recoverable XML error while parsing document: %s", err->message)
+        }
+    }
+
     mce_start_children(reader) {
         mce_start_element(reader, NULL, _X("t")) {
             mce_skip_attributes(reader);
@@ -18,10 +28,14 @@ void dump_text(mceTextReader_t *reader, dyn_buffer_t *buf) {
         } mce_end_element(reader);
 
         mce_start_element(reader, NULL, NULL) {
-            dump_text(reader, buf);
+            int ret = dump_text(reader, buf);
+            if (ret != 0) {
+                return ret;
+            }
         } mce_end_element(reader);
 
     } mce_end_children(reader)
+    return 0;
 }
 
 __always_inline
@@ -52,23 +66,28 @@ int should_read_part(opcPart part) {
 }
 
 __always_inline
-void read_part(opcContainer *c, dyn_buffer_t *buf, opcPart part, document_t *doc) {
+int 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 ret = opcXmlReaderOpen(c, &reader, part, NULL, "UTF-8", XML_PARSE_NOWARNING | XML_PARSE_NOERROR | XML_PARSE_NONET);
 
     if (ret != OPC_ERROR_NONE) {
         LOG_ERRORF(doc->filepath, "(doc.c) opcXmlReaderOpen() returned error code %d", ret);
-        return;
+        return -1;
     }
 
     mce_start_document(&reader) {
         mce_start_element(&reader, NULL, NULL) {
-                dump_text(&reader, buf);
+            ret = dump_text(&reader, buf);
+            if (ret != 0) {
+                mceTextReaderCleanup(&reader);
+                return -1;
+            }
         } mce_end_element(&reader);
     } mce_end_document(&reader);
 
     mceTextReaderCleanup(&reader);
+    return 0;
 }
 
 void parse_doc(void *mem, size_t mem_len, document_t *doc) {
@@ -88,7 +107,10 @@ void parse_doc(void *mem, size_t mem_len, document_t *doc) {
     opcPart part = opcPartGetFirst(c);
     do {
         if (should_read_part(part)) {
-            read_part(c, &buf, part, doc);
+            int ret = read_part(c, &buf, part, doc);
+            if (ret != 0) {
+                break;
+            }
         }
     } while ((part = opcPartGetNext(c, part)));
 

src/parsing/media.c

@@ -193,6 +193,24 @@ 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)
+            } else if (strcmp(tag->key, "DateTime") == 0) {
+                APPEND_TAG_META(doc, tag, MetaExifDateTime)
             }
         }
     }

src/parsing/pdf.c

@@ -141,7 +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);
@@ -157,8 +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/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>

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 +31,16 @@ 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,
+    MetaExifDateTime = 23 | 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

@@ -7,7 +7,7 @@
 #define INITIAL_BUF_SIZE 1024 * 16
 
 #define SHOULD_IGNORE_CHAR(c) !(SHOULD_KEEP_CHAR(c))
-#define SHOULD_KEEP_CHAR(c) ((c >= '0' && c <= '9') || (c >= 'A' && c <= 'z') || (c > 127))
+#define SHOULD_KEEP_CHAR(c) ((c >= '\'' && c <= ';') || (c >= 'A' && c <= 'z') || (c > 127))
 
 
 typedef struct dyn_buffer {
@@ -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

@@ -221,14 +221,6 @@ int search(UNUSED(void *p), onion_request *req, onion_response *res) {
         return OCS_NOT_PROCESSED;
     }
 
-    char *scroll_param;
-    const char *scroll = onion_request_get_query(req, "scroll");
-    if (scroll != NULL) {
-        scroll_param = "?scroll=3m";
-    } else {
-        scroll_param = "";
-    }
-
     const struct onion_block_t *block = onion_request_get_data(req);
 
     if (block == NULL) {
@@ -236,7 +228,7 @@ int search(UNUSED(void *p), onion_request *req, onion_response *res) {
     }
 
     char url[4096];
-    snprintf(url, 4096, "%s/sist2/_search%s", WebCtx.es_url, scroll_param);
+    snprintf(url, 4096, "%s/sist2/_search", WebCtx.es_url);
     response_t *r = web_post(url, onion_block_data(block), "Content-Type: application/json");
 
     set_default_headers(res);
@@ -254,43 +246,6 @@ int search(UNUSED(void *p), onion_request *req, onion_response *res) {
     return OCS_PROCESSED;
 }
 
-int scroll(UNUSED(void *p), onion_request *req, onion_response *res) {
-
-    int flags = onion_request_get_flags(req);
-    if ((flags & OR_METHODS) != OR_GET) {
-        return OCS_NOT_PROCESSED;
-    }
-
-    char url[4096];
-    snprintf(url, 4096, "%s/_search/scroll", WebCtx.es_url);
-
-    const char *scroll_id = onion_request_get_query(req, "scroll_id");
-
-    cJSON *json = cJSON_CreateObject();
-    cJSON_AddStringToObject(json, "scroll_id", scroll_id);
-    cJSON_AddStringToObject(json, "scroll", "3m");
-
-    char *json_str = cJSON_PrintUnformatted(json);
-    response_t *r = web_post(url, json_str, "Content-Type: application/json");
-
-    cJSON_Delete(json);
-    cJSON_free(json_str);
-
-    if (r->status_code != 200) {
-        free_response(r);
-        return OCS_NOT_PROCESSED;
-    }
-
-    set_default_headers(res);
-    onion_response_set_header(res, "Content-Type", "application/json");
-    onion_response_set_header(res, "Content-Disposition", "application/json");
-    onion_response_set_length(res, r->size);
-    onion_response_write(res, r->body, r->size);
-    free_response(r);
-
-    return OCS_PROCESSED;
-}
-
 int serve_file_from_url(cJSON *json, index_t *idx, onion_request *req, onion_response *res) {
 
     const char *path = cJSON_GetObjectItem(json, "path")->valuestring;
@@ -431,6 +386,23 @@ int file(UNUSED(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);
@@ -449,7 +421,7 @@ void serve(const char *hostname, const char *port) {
     onion_url_add(urls, "img/sprite-skin-flat.png", img_sprite_skin_flag);
 
     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})/"

web/js/dom.js

@@ -113,7 +113,7 @@ function getTags(hit, mimeCategory) {
     switch (mimeCategory) {
         case "video":
         case "image":
-            if (hit["_source"].hasOwnProperty("videoc")) {
+            if (hit["_source"].hasOwnProperty("videoc") && hit["_source"]["videoc"]) {
                 const formatTag = document.createElement("span");
                 formatTag.setAttribute("class", "badge badge-pill badge-video");
                 formatTag.appendChild(document.createTextNode(hit["_source"]["videoc"].replace(" ", "")));
@@ -121,7 +121,7 @@ function getTags(hit, mimeCategory) {
             }
             break;
         case "audio": {
-            if (hit["_source"].hasOwnProperty("audioc")) {
+            if (hit["_source"].hasOwnProperty("audioc") && hit["_source"]["audioc"]) {
                 let formatTag = document.createElement("span");
                 formatTag.setAttribute("class", "badge badge-pill badge-audio");
                 formatTag.appendChild(document.createTextNode(hit["_source"]["audioc"]));
@@ -499,8 +499,7 @@ function makePreloader() {
 function makePageIndicator(searchResult) {
     let pageIndicator = document.createElement("div");
     pageIndicator.setAttribute("class", "page-indicator font-weight-light");
-    const totalHits = searchResult["hits"]["total"].hasOwnProperty("value")
-        ? searchResult["hits"]["total"]["value"] : searchResult["hits"]["total"];
+    const totalHits = searchResult["aggregations"]["total_count"]["value"];
     pageIndicator.appendChild(document.createTextNode(docCount + " / " + totalHits));
     return pageIndicator;
 }
@@ -547,8 +546,7 @@ function makeStatsCard(searchResult) {
     });
 
     let stat = document.createElement("span");
-    const totalHits = searchResult["hits"]["total"].hasOwnProperty("value")
-        ? searchResult["hits"]["total"]["value"] : searchResult["hits"]["total"];
+    const totalHits = searchResult["aggregations"]["total_count"]["value"];
     stat.appendChild(document.createTextNode(totalHits + " results in " + searchResult["took"] + "ms"));
 
     statsCardBody.appendChild(stat);

web/js/search.js

@@ -6,7 +6,8 @@ let tagTree;
 
 let searchBar = document.getElementById("searchBar");
 let pathBar = document.getElementById("pathBar");
-let scroll_id = null;
+let lastDoc = null;
+let reachedEnd = false;
 let docCount = 0;
 let coolingDown = false;
 let searchBusy = true;
@@ -259,41 +260,18 @@ function insertHits(resultContainer, hits) {
 }
 
 window.addEventListener("scroll", function () {
-    if (!coolingDown && !searchBusy) {
+    if (!searchBusy) {
         let threshold = 400;
 
         if ((window.innerHeight + window.scrollY) >= document.body.offsetHeight - threshold) {
-            coolingDown = true;
-            doScroll();
+            if (!reachedEnd) {
+                coolingDown = true;
+                search(lastDoc);
+            }
         }
     }
 });
 
-function doScroll() {
-    $.get("scroll", {scroll_id: scroll_id})
-        .then(searchResult => {
-            let searchResults = document.getElementById("searchResults");
-            let hits = searchResult["hits"]["hits"];
-
-            //Page indicator
-            let pageIndicator = makePageIndicator(searchResult);
-            searchResults.appendChild(pageIndicator);
-
-            //Result container
-            let resultContainer = makeResultContainer();
-            searchResults.appendChild(resultContainer);
-
-            insertHits(resultContainer, hits);
-
-            if (hits.length === SIZE) {
-                coolingDown = false;
-            }
-        })
-        .fail(() => {
-            window.location.reload();
-        })
-}
-
 function getSelectedNodes(tree) {
     let selectedNodes = [];
 
@@ -314,21 +292,25 @@ function getSelectedNodes(tree) {
     return selectedNodes
 }
 
-function search() {
+function search(after = null) {
+    lastDoc = null;
+
     if (searchBusy) {
         return;
     }
     searchBusy = true;
 
-    //Clear old search results
     let searchResults = document.getElementById("searchResults");
-    while (searchResults.firstChild) {
-        searchResults.removeChild(searchResults.firstChild);
+    //Clear old search results
+    let preload;
+    if (!after) {
+        while (searchResults.firstChild) {
+            searchResults.removeChild(searchResults.firstChild);
+        }
+        preload = makePreloader();
+        searchResults.appendChild(preload);
     }
 
-    const preload = makePreloader();
-    searchResults.appendChild(preload);
-
     let query = searchBar.value;
     let empty = query === "";
     let condition = empty ? "should" : "must";
@@ -362,9 +344,9 @@ function search() {
         filters.push([{terms: {"tag": tags}}]);
     }
 
-    $.jsonPost("es?scroll=1", {
+    let q = {
         "_source": {
-            excludes: ["content"]
+            excludes: ["content", "_tie"]
         },
         query: {
             bool: {
@@ -379,8 +361,9 @@ function search() {
                 filter: filters
             }
         },
-        sort: [
-            "_score"
+        "sort": [
+            {"_score": {"order": "desc"}},
+            {"_tie": {"order": "asc"}}
         ],
         highlight: {
             pre_tags: ["<mark>"],
@@ -393,24 +376,41 @@ function search() {
                 font_name: {},
             }
         },
-        aggs: {
-            total_size: {"sum": {"field": "size"}}
-        },
+        aggs:
+            {
+                total_size: {"sum": {"field": "size"}},
+                total_count: {"value_count": {"field": "size"}}
+            },
         size: SIZE,
-    }).then(searchResult => {
-        scroll_id = searchResult["_scroll_id"];
+    };
 
-        preload.remove();
-        //Search stats
-        searchResults.appendChild(makeStatsCard(searchResult));
+    if (after) {
+        q.search_after = [after["_score"], after["_id"]];
+    }
+
+    $.jsonPost("es", q).then(searchResult => {
+        let hits = searchResult["hits"]["hits"];
+        if (hits) {
+            lastDoc = hits[hits.length - 1];
+        }
+
+        if (!after) {
+            preload.remove();
+            searchResults.appendChild(makeStatsCard(searchResult));
+        } else {
+            let pageIndicator = makePageIndicator(searchResult);
+            searchResults.appendChild(pageIndicator);
+        }
 
         //Setup page
         let resultContainer = makeResultContainer();
         searchResults.appendChild(resultContainer);
 
-        docCount = 0;
-        insertHits(resultContainer, searchResult["hits"]["hits"]);
-
+        if (!after) {
+            docCount = 0;
+        }
+        reachedEnd = hits.length !== SIZE;
+        insertHits(resultContainer, hits);
         searchBusy = false;
     });
 }

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">v1.2.3</span>
+    <span class="badge badge-pill version">v1.2.9</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>