Handle XML errors #18

Docker/build.sh

@@ -5,8 +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

@@ -95,7 +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",
@@ -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;"
+      }
+    }
+  ]
+}

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:
+    with open(file, "r") as f:
-        data = f.read()
+        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/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);

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 = "1.2.4";
+static const char *const Version = "1.2.7";
 static const char *const usage[] = {
         "sist2 scan [OPTION]... PATH",
         "sist2 index [OPTION]... INDEX",

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
@@ -52,30 +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 options;
+    int ret = opcXmlReaderOpen(c, &reader, part, NULL, "UTF-8", XML_PARSE_NOWARNING | XML_PARSE_NOERROR | XML_PARSE_NONET);
-    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);
-        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) {
@@ -95,7 +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 +193,22 @@ 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/types.h

@@ -2,9 +2,9 @@
 #define SIST2_TYPES_H
 
 
-#define META_INT_MASK 0xF0
+#define META_INT_MASK 0x80
-#define META_STR_MASK 0xE0
+#define META_STR_MASK 0x40
-#define META_LONG_MASK 0xD0
+#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

@@ -91,7 +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
@@ -172,8 +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 (SHOULD_IGNORE_CHAR(c) || c == ' ') {
-        if (!buf->last_char_was_whitespace) {
+        if (!buf->last_char_was_whitespace && buf->dyn_buffer.cur != 0) {
             dyn_buffer_write_char(&buf->dyn_buffer, ' ');
             buf->last_char_was_whitespace = TRUE;
 

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;
@@ -466,7 +421,6 @@ 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,

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"]));

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;
+            if (!reachedEnd) {
-            doScroll();
+                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) {
+    //Clear old search results
-        searchResults.removeChild(searchResults.firstChild);
+    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: [
+        "sort": [
-            "_score"
+            {"_score": {"order": "desc"}},
+            {"_tie": {"order":"asc"}}
         ],
         highlight: {
             pre_tags: ["<mark>"],
@@ -397,20 +380,35 @@ function search() {
             total_size: {"sum": {"field": "size"}}
         },
         size: SIZE,
-    }).then(searchResult => {
+    }
-        scroll_id = searchResult["_scroll_id"];
 
-        preload.remove();
+    if (after) {
-        //Search stats
+        q.search_after = [after["_score"], after["_id"]];
-        searchResults.appendChild(makeStatsCard(searchResult));
+    }
+
+    $.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;
+        if (!after) {
-        insertHits(resultContainer, searchResult["hits"]["hits"]);
+            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">v1.2.4</span>
+    <span class="badge badge-pill version">v1.2.7</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>