Manual tagging

README.md

@@ -18,7 +18,7 @@ sist2 (Simple incremental search tool)
 * Extracts text and metadata from common file types \*
 * Generates thumbnails \*
 * Incremental scanning
-* Automatic tagging from file attributes via [user scripts](docs/scripting.md)
+* Manual tagging from the UI and automatic tagging based on file attributes via [user scripts](docs/scripting.md)
 * Recursive scan inside archive files \*\*
 * OCR support with tesseract \*\*\*
 * Stats page & disk utilisation visualization

docs/USAGE.md

@@ -15,6 +15,7 @@
     * [rewrite_url](#rewrite_url)
     * [link to specific indices](#link-to-specific-indices)
 * [exec-script](#exec-script)
+* [tagging](#tagging)
 
 ```
 Usage: sist2 scan [OPTION]... PATH
@@ -57,6 +58,7 @@ Web options
     --es-url=<str>                Elasticsearch url. DEFAULT=http://localhost:9200
     --bind=<str>                  Listen on this address. DEFAULT=localhost:4090
     --auth=<str>                  Basic auth in user:password format
+    --tag-auth=<str>              Basic auth in user:password format for tagging
 
 Exec-script options
     --script-file=<str>           Path to user script.
@@ -145,7 +147,10 @@ documents.idx/
 ├── agg_mime.csv
 ├── agg_date.csv
 ├── add_size.csv
-└── thumbs
+├── thumbs
+|   ├── data.mdb
+|   └── lock.mdb
+└── tags
     ├── data.mdb
     └── lock.mdb
 ```
@@ -270,6 +275,8 @@ sist2 index --print ./my_index/ | jq | less
  * `--es-url=<str>` Elasticsearch url.
  * `--bind=<str>` Listen on this address.
  * `--auth=<str>` Basic auth in user:password format
+ * `--tag-auth=<str>` Basic auth in user:password format. Works the same way as the 
+    `--auth` argument, but authentication is only applied the `/tag/` endpoint.
  
 ### Web examples
 
@@ -301,3 +308,31 @@ not displayed.
 ## exec-script
 
 The `exec-script` command is used to execute a user script for an index that has already been imported to Elasticsearch with the `index` command. Note that the documents will not be reset to their default state before each execution as the `index` command does: if you make undesired changes to the documents by accident, you will need to run `index` again to revert to the original state.
+
+
+# Tagging
+
+### Manual tagging
+
+You can modify tags of individual documents directly from the 
+ `web` interface. Note that you can setup authentication for this feature
+ with the `--tag-auth` option (See [web options](#web-options))
+
+![manual_tag](docs/manual_tag.png)
+
+Tags that are manually added are saved both in the 
+ index folder (in `/tags/`) and in Elasticsearch*. When re-`index`ing, 
+ they are read from the index and automatically applied.
+ 
+You can safely copy the `/tags/` database to another index.
+
+See [Automatic tagging](#automatic-tagging) for information about tag 
+ hierarchies and tag colors.
+
+\* *It can take a few seconds to take effect in new search queries, and the page needs 
+    to be reloaded for the tag tab to update*
+
+
+### Automatic tagging
+
+See [scripting](docs/scripting.md) documentation.

src/cli.c

@@ -326,7 +326,7 @@ int web_args_validate(web_args_t *args, int argc, const char **argv) {
         }
 
         strncpy(args->auth_user, args->credentials, (ptr - args->credentials));
-        strcpy(args->auth_pass, ptr);
+        strcpy(args->auth_pass, ptr + 1);
 
         if (strlen(args->auth_user) == 0) {
             fprintf(stderr, "--auth username must be at least one character long");
@@ -338,6 +338,31 @@ int web_args_validate(web_args_t *args, int argc, const char **argv) {
         args->auth_enabled = FALSE;
     }
 
+    if (args->tag_credentials != NULL && args->credentials != NULL) {
+        fprintf(stderr, "--auth and --tag-auth are mutually exclusive");
+        return 1;
+    }
+
+    if (args->tag_credentials != NULL) {
+        char *ptr = strstr(args->tag_credentials, ":");
+        if (ptr == NULL) {
+            fprintf(stderr, "Invalid --tag-auth format, see usage\n");
+            return 1;
+        }
+
+        strncpy(args->auth_user, args->tag_credentials, (ptr - args->tag_credentials));
+        strcpy(args->auth_pass, ptr + 1);
+
+        if (strlen(args->auth_user) == 0) {
+            fprintf(stderr, "--tag-auth username must be at least one character long");
+            return 1;
+        }
+
+        args->tag_auth_enabled = TRUE;
+    } else {
+        args->tag_auth_enabled = FALSE;
+    }
+
     args->index_count = argc - 1;
     args->indices = argv + 1;
 
@@ -352,6 +377,7 @@ int web_args_validate(web_args_t *args, int argc, const char **argv) {
     LOG_DEBUGF("cli.c", "arg es_url=%s", args->es_url)
     LOG_DEBUGF("cli.c", "arg listen=%s", args->listen_address)
     LOG_DEBUGF("cli.c", "arg credentials=%s", args->credentials)
+    LOG_DEBUGF("cli.c", "arg tag_credentials=%s", args->tag_credentials)
     LOG_DEBUGF("cli.c", "arg auth_user=%s", args->auth_user)
     LOG_DEBUGF("cli.c", "arg auth_pass=%s", args->auth_pass)
     LOG_DEBUGF("cli.c", "arg index_count=%d", args->index_count)

src/cli.h

@@ -48,9 +48,11 @@ typedef struct web_args {
     char *es_url;
     char *listen_address;
     char *credentials;
+    char *tag_credentials;
     char auth_user[256];
     char auth_pass[256];
     int auth_enabled;
+    int tag_auth_enabled;
     int index_count;
     const char **indices;
 } web_args_t;

src/ctx.h

@@ -60,6 +60,8 @@ typedef struct {
     char *es_url;
     int batch_size;
     tpool_t *pool;
+    store_t *tag_store;
+    GHashTable *tags;
 } IndexCtx_t;
 
 typedef struct {
@@ -68,6 +70,7 @@ typedef struct {
     char *auth_user;
     char *auth_pass;
     int auth_enabled;
+    int tag_auth_enabled;
     struct index_t indices[64];
 } WebCtx_t;
 

src/io/serialize.c

@@ -62,7 +62,7 @@ index_descriptor_t read_index_descriptor(char *path) {
     int fd = open(path, O_RDONLY);
 
     if (fd == -1) {
-        LOG_FATALF("serialize.c", "Invalid/corrupt index (Could not find descriptor): %s: %s\n", path ,strerror(errno))
+        LOG_FATALF("serialize.c", "Invalid/corrupt index (Could not find descriptor): %s: %s\n", path, strerror(errno))
     }
 
     char *buf = malloc(info.st_size + 1);
@@ -172,8 +172,8 @@ void write_document(document_t *doc) {
     dyn_buffer_t buf = dyn_buffer_create();
 
     // Ignore root directory in the file path
-    doc->ext = doc->ext - ScanCtx.index.desc.root_len;
+    doc->ext = (short) (doc->ext - ScanCtx.index.desc.root_len);
-    doc->base = doc->base - ScanCtx.index.desc.root_len;
+    doc->base = (short) (doc->base - ScanCtx.index.desc.root_len);
     doc->filepath += ScanCtx.index.desc.root_len;
 
     dyn_buffer_write(&buf, doc, sizeof(line_t));
@@ -230,7 +230,7 @@ void read_index_bin(const char *path, const char *index_id, index_func func) {
         char uuid_str[UUID_STR_LEN];
         uuid_unparse(line.uuid, uuid_str);
 
-        const char* mime_text = mime_get_mime_text(line.mime);
+        const char *mime_text = mime_get_mime_text(line.mime);
         if (mime_text == NULL) {
             cJSON_AddNullToObject(document, "mime");
         } else {
@@ -239,12 +239,18 @@ void read_index_bin(const char *path, const char *index_id, index_func func) {
         cJSON_AddNumberToObject(document, "size", (double) line.size);
         cJSON_AddNumberToObject(document, "mtime", line.mtime);
 
-        int c;
+        int c = 0;
         while ((c = getc(file)) != 0) {
             dyn_buffer_write_char(&buf, (char) c);
         }
         dyn_buffer_write_char(&buf, '\0');
 
+        const char *tags_string = g_hash_table_lookup(IndexCtx.tags, buf.buf);
+        if (tags_string != NULL) {
+            cJSON *tags_arr = cJSON_Parse(tags_string);
+            cJSON_AddItemToObject(document, "tag", tags_arr);
+        }
+
         cJSON_AddStringToObject(document, "extension", buf.buf + line.ext);
         if (*(buf.buf + line.ext - 1) == '.') {
             *(buf.buf + line.ext - 1) = '\0';

src/io/store.c

@@ -111,3 +111,35 @@ char *store_read(store_t *store, char *key, size_t key_len, size_t *ret_vallen)
     return buf;
 }
 
+GHashTable *store_read_all(store_t *store) {
+
+    int count = 0;
+
+    GHashTable *table = g_hash_table_new_full(g_str_hash, g_str_equal, free, free);
+
+    MDB_txn *txn = NULL;
+    mdb_txn_begin(store->env, NULL, MDB_RDONLY, &txn);
+
+    MDB_cursor *cur = NULL;
+    mdb_cursor_open(txn, store->dbi, &cur);
+
+    MDB_val key;
+    MDB_val value;
+
+    while (mdb_cursor_get(cur, &key, &value, MDB_NEXT) == 0) {
+        char *key_str = malloc(key.mv_size);
+        memcpy(key_str, key.mv_data, key.mv_size);
+        char *val_str = malloc(value.mv_size);
+        memcpy(val_str, value.mv_data, value.mv_size);
+
+        g_hash_table_insert(table, key_str, val_str);
+        count += 1;
+    }
+
+    LOG_DEBUGF("store.c", "Read tags for %d documents", count);
+
+    mdb_cursor_close(cur);
+    mdb_txn_abort(txn);
+    return table;
+}
+

src/io/store.h

@@ -4,6 +4,8 @@
 #include <pthread.h>
 #include <lmdb.h>
 
+#include <glib.h>
+
 #define STORE_SIZE_TN 1024 * 1024 * 5
 #define STORE_SIZE_TAG 1024 * 16
 
@@ -23,4 +25,6 @@ void store_write(store_t *store, char *key, size_t key_len, char *buf, size_t bu
 
 char *store_read(store_t *store, char *key, size_t key_len, size_t *ret_vallen);
 
+GHashTable *store_read_all(store_t *store);
+
 #endif

src/main.c

@@ -21,7 +21,7 @@
 #define EPILOG "Made by simon987 <me@simon987.net>. Released under GPL-3.0"
 
 
-static const char *const Version = "2.5.2";
+static const char *const Version = "2.6.0";
 static const char *const usage[] = {
         "sist2 scan [OPTION]... PATH",
         "sist2 index [OPTION]... INDEX",
@@ -271,6 +271,12 @@ void sist2_index(index_args_t *args) {
         LOG_FATALF("main.c", "Could not open index %s: %s", args->index_path, strerror(errno))
     }
 
+    char path_tmp[PATH_MAX];
+    snprintf(path_tmp, sizeof(path_tmp), "%s/tags", args->index_path);
+    mkdir(path_tmp, S_IWUSR | S_IRUSR | S_IXUSR);
+    IndexCtx.tag_store = store_create(path_tmp, STORE_SIZE_TAG);
+    IndexCtx.tags = store_read_all(IndexCtx.tag_store);
+
     index_func f;
     if (args->print) {
         f = print_json;
@@ -303,8 +309,11 @@ void sist2_index(index_args_t *args) {
     if (!args->print) {
         finish_indexer(args->script, desc.uuid);
     }
-
     tpool_destroy(IndexCtx.pool);
+
+    store_destroy(IndexCtx.tag_store);
+    g_hash_table_remove_all(IndexCtx.tags);
+    g_hash_table_destroy(IndexCtx.tags);
 }
 
 void sist2_exec_script(exec_args_t *args) {
@@ -330,6 +339,7 @@ void sist2_web(web_args_t *args) {
     WebCtx.auth_user = args->auth_user;
     WebCtx.auth_pass = args->auth_pass;
     WebCtx.auth_enabled = args->auth_enabled;
+    WebCtx.tag_auth_enabled = args->tag_auth_enabled;
 
     for (int i = 0; i < args->index_count; i++) {
         char *abs_path = abspath(args->indices[i]);
@@ -419,6 +429,7 @@ int main(int argc, const char *argv[]) {
             OPT_STRING(0, "es-url", &common_es_url, "Elasticsearch url. DEFAULT=http://localhost:9200"),
             OPT_STRING(0, "bind", &web_args->listen_address, "Listen on this address. DEFAULT=localhost:4090"),
             OPT_STRING(0, "auth", &web_args->credentials, "Basic auth in user:password format"),
+            OPT_STRING(0, "tag-auth", &web_args->tag_credentials, "Basic auth in user:password format for tagging"),
 
             OPT_GROUP("Exec-script options"),
             OPT_STRING(0, "script-file", &common_script_path, "Path to user script."),

src/static/css/dark.css

@@ -197,6 +197,18 @@ a:hover,.btn:hover {
     margin-right: 3px;
 }
 
+.badge-delete {
+    margin-right: -2px;
+    margin-left: 2px;
+    margin-top: -1px;
+    font-family: monospace;
+    font-size: 90%;
+    background: rgba(0,0,0,0.2);
+    padding: 0.1em 0.4em;
+    color: white;
+    cursor: pointer;
+}
+
 .badge-user {
     color: #212529;
     background-color: #e0e0e0;

src/static/css/light.css

@@ -113,6 +113,7 @@ body {
 .badge-delete {
     margin-right: -2px;
     margin-left: 2px;
+    margin-top: -1px;
     font-family: monospace;
     font-size: 90%;
     background: rgba(0,0,0,0.2);

src/static/js/search.js

@@ -117,7 +117,7 @@ window.onload = () => {
         minChars: 1,
         delay: 200,
         renderItem: function (item) {
-            return '<div class="autocomplete-suggestion" data-val="' + item + '">' + item + '</div>';
+            return '<div class="autocomplete-suggestion" data-val="' + item + '">' + item.split("#")[0] + '</div>';
         },
         source: async function (term, suggest) {
             term = term.toLowerCase();
@@ -132,12 +132,18 @@ window.onload = () => {
             }
             suggest(matches.sort());
         },
-        onSelect: function () {
+        onSelect: function (e, item) {
+            const name = item.split("#")[0];
+            const color = "#" + item.split("#")[1];
+            $("#tag-color").val(color);
+            $("#tag-color").trigger("keyup", color);
+            tagBar.value = name;
+            e.preventDefault();
         }
     });
     [tagBar, document.getElementById("tag-color")].forEach(elem => {
         elem.addEventListener("keyup", e => {
-            if (e.key === "Enter") {
+            if (e.key === "Enter" && tagBar.value.length > 0) {
                 const tag = tagBar.value + document.getElementById("tag-color").value;
                 saveTag(tag, currentDocToTag).then(() => currentTagCallback(tag));
             }
@@ -162,7 +168,7 @@ window.onload = () => {
 };
 
 function saveTag(tag, hit) {
-    const relPath = hit["_source"]["path"] + "/" + hit["_source"] + ext(hit);
+    const relPath = hit["_source"]["path"] + "/" + hit["_source"]["name"] + ext(hit);
 
     return $.jsonPost("/tag/" + hit["_source"]["index"], {
         delete: false,
@@ -186,7 +192,7 @@ function saveTag(tag, hit) {
 }
 
 function deleteTag(tag, hit) {
-    const relPath = hit["_source"]["path"] + "/" + hit["_source"] + ext(hit);
+    const relPath = hit["_source"]["path"] + "/" + hit["_source"]["name"] + ext(hit);
 
     return $.jsonPost("/tag/" + hit["_source"]["index"], {
         delete: true,
@@ -344,11 +350,14 @@ $.jsonPost("es", {
 });
 
 function addTag(map, tag, id, count) {
-    let tags = tag.split("#")[0].split(".");
+    // let tags = tag.split("#")[0].split(".");
+    let tags = tag.split(".");
 
     let child = {
         id: id,
-        text: tags.length !== 1 ? tags[0] : `${tags[0]} (${count})`,
+        values: [id],
+        count: count,
+        text: tags.length !== 1 ? tags[0] : `${tags[0].split("#")[0]} (${count})`,
         name: tags[0],
         children: [],
         isLeaf: tags.length === 1,
@@ -396,10 +405,15 @@ function addTag(map, tag, id, count) {
 
     let found = false;
     map.forEach(node => {
-        if (node.name === child.name) {
+        if (node.name.split("#")[0] === child.name.split("#")[0]) {
             found = true;
             if (tags.length !== 1) {
                 addTag(node.children, tags.slice(1).join("."), id, count);
+            } else {
+                // Same name, different color
+                node.count += count;
+                node.text = `${tags[0].split("#")[0]} (${node.count})`;
+                node.values.push(id);
             }
         }
     });
@@ -451,7 +465,11 @@ function getSelectedNodes(tree) {
 
         //Only get children
         if (selected[i].text.indexOf("(") !== -1) {
-            selectedNodes.push(selected[i].id);
+            if (selected[i].values) {
+                selectedNodes.push(selected[i].values);
+            } else {
+                selectedNodes.push(selected[i].id);
+            }
         }
     }
 
@@ -514,7 +532,9 @@ function search(after = null) {
 
     let tags = getSelectedNodes(tagTree);
     if (!tags.includes("any")) {
-        tags.forEach(term => filters.push({term: {"tag": term}}))
+        tags.forEach(tagGroup => {
+            filters.push({terms: {"tag": tagGroup}})
+        })
     }
 
     if (date_min && date_max) {
@@ -758,6 +778,7 @@ function getNextDepth(node) {
                     text: `${name}/ (${bucket.doc_count})`,
                     depth: node.depth + 1,
                     index: node.index,
+                    values: [bucket.key],
                     children: true,
                 }
             }).filter(x => x !== null)
@@ -788,6 +809,7 @@ function createPathTree(target) {
     selectedIndices.forEach(index => {
         pathTree.addNode({
             id: "/" + index,
+            values: ["/" + index],
             text: `/[${indexMap[index]}]`,
             index: index,
             depth: 0,
@@ -838,8 +860,8 @@ function getTagChoices() {
             resp["suggest"]["tag"][0]["options"].map(opt => opt["_source"]["tag"]).forEach(tags => {
                 tags.forEach(tag => {
                     const t = tag.split("#")[0];
-                    if (result.indexOf(t) === -1) {
+                    if (!result.find(x => x.split("#")[0] === t)) {
-                        result.push(t);
+                        result.push(tag);
                     }
                 });
             });

src/static/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">2.5.2</span>
+    <span class="badge badge-pill version">2.6.0</span>
     <span class="tagline">Lightning-fast file system indexer and search tool </span>
     <a class="btn ml-auto" href="/stats">Stats</a>
     <button class="btn" type="button" data-toggle="modal" data-target="#settings" onclick="loadSettings()">Settings

src/static/stats.html

@@ -10,7 +10,7 @@
 
 <nav class="navbar navbar-expand-lg">
     <a class="navbar-brand" href="/">sist2</a>
-    <span class="badge badge-pill version">2.5.2</span>
+    <span class="badge badge-pill version">2.6.0</span>
     <span class="tagline">Lightning-fast file system indexer and search tool </span>
     <a style="margin-left: auto" class="btn" href="/">Back</a>
     <button class="btn" type="button" data-toggle="modal" data-target="#settings"

src/web/serve.c

@@ -562,7 +562,7 @@ void tag(struct mg_connection *nc, struct http_message *hm, struct mg_str *path)
         snprintf(buf, sizeof(buf),
                  "{"
                  "    \"script\" : {"
-                 "        \"source\": \"ctx._source.tag.add(params.tag)\","
+                 "        \"source\": \"if(ctx._source.tag == null) {ctx._source.tag = new ArrayList()} ctx._source.tag.add(params.tag)\","
                  "        \"lang\": \"painless\","
                  "        \"params\" : {"
                  "            \"tag\" : \"%s\""
@@ -577,7 +577,7 @@ void tag(struct mg_connection *nc, struct http_message *hm, struct mg_str *path)
     }
 
     char *json_str = cJSON_PrintUnformatted(arr);
-    store_write(store, arg_req->relpath, strlen(arg_req->relpath), json_str, strlen(json_str) + 1);
+    store_write(store, arg_req->relpath, strlen(arg_req->relpath) + 1, json_str, strlen(json_str) + 1);
 
     free(arg_req);
     free(json_str);
@@ -586,6 +586,21 @@ void tag(struct mg_connection *nc, struct http_message *hm, struct mg_str *path)
     free(body);
 }
 
+int validate_auth(struct mg_connection *nc, struct http_message *hm) {
+    char user[256] = {0,};
+    char pass[256] = {0,};
+
+    int ret = mg_get_http_basic_auth(hm, user, sizeof(user), pass, sizeof(pass));
+    if (ret == -1 || strcmp(user, WebCtx.auth_user) != 0 || strcmp(pass, WebCtx.auth_pass) != 0) {
+        mg_printf(nc, "HTTP/1.1 401 Unauthorized\r\n"
+                      "WWW-Authenticate: Basic realm=\"sist2\"\r\n"
+                      "Content-Length: 0\r\n\r\n");
+        nc->flags |= MG_F_SEND_AND_CLOSE;
+        return FALSE;
+    }
+    return TRUE;
+}
+
 static void ev_router(struct mg_connection *nc, int ev, void *p) {
     struct mg_str scheme;
     struct mg_str user_info;
@@ -606,15 +621,7 @@ static void ev_router(struct mg_connection *nc, int ev, void *p) {
 
 
         if (WebCtx.auth_enabled == TRUE) {
-            char user[256] = {0,};
+            if (!validate_auth(nc, hm)) {
-            char pass[256] = {0,};
-
-            int ret = mg_get_http_basic_auth(hm, user, sizeof(user), pass, sizeof(pass));
-            if (ret == -1 || strcmp(user, WebCtx.auth_user) != 0 || strcmp(pass, WebCtx.auth_pass) != 0) {
-                mg_printf(nc, "HTTP/1.1 401 Unauthorized\r\n"
-                              "WWW-Authenticate: Basic realm=\"sist2\"\r\n"
-                              "Content-Length: 0\r\n\r\n");
-                nc->flags |= MG_F_SEND_AND_CLOSE;
                 return;
             }
         }
@@ -644,6 +651,11 @@ static void ev_router(struct mg_connection *nc, int ev, void *p) {
         } else if (has_prefix(&path, &((struct mg_str) MG_MK_STR("/s/")))) {
             stats_files(nc, hm, &path);
         } else if (has_prefix(&path, &((struct mg_str) MG_MK_STR("/tag/")))) {
+            if (WebCtx.tag_auth_enabled == TRUE) {
+                if (!validate_auth(nc, hm)) {
+                    return;
+                }
+            }
             tag(nc, hm, &path);
         } else if (has_prefix(&path, &((struct mg_str) MG_MK_STR("/d/")))) {
             document_info(nc, hm, &path);