initial changelog and container ES schemas

2 files changed, 113 insertions, 0 deletions

diff --git a/extra/elasticsearch/changelog_schema.json b/extra/elasticsearch/changelog_schema.json
new file mode 100644
index 00000000..7a7ec90c
--- /dev/null
+++ b/extra/elasticsearch/changelog_schema.json
@@ -0,0 +1,39 @@
+{
+"settings": {
+    "index": {
+        "analysis": {
+            "analyzer": {
+                "default": {
+                    "type": "custom",
+                    "tokenizer": "standard",
+                    "filter": [ "lowercase", "asciifolding" ]
+                }
+            }
+        }
+    }
+},
+"mappings": {
+    "changelog": {
+        "properties": {
+            "index":            { "type": "integer" },
+            "editgorup_id":     { "type": "keyword" },
+            "timestamp":        { "type": "datetime" },
+            "username":         { "type": "keyword" },
+            "is_bot":           { "type": "boolean" },
+            "is_admin":         { "type": "boolean" },
+            "agent":            { "type": "keyword" },
+            "containers":       { "type": "integer" },
+            "creators":         { "type": "integer" },
+            "files":            { "type": "integer" },
+            "filessets":        { "type": "integer" },
+            "webcaptures":      { "type": "integer" },
+            "releases":         { "type": "integer" },
+            "works":            { "type": "integer" },
+            "created":          { "type": "integer" },
+            "updated":          { "type": "integer" },
+            "deleted":          { "type": "integer" },
+            "total":            { "type": "integer" }
+        }
+    }
+}
+}
diff --git a/extra/elasticsearch/container_schema.json b/extra/elasticsearch/container_schema.json
new file mode 100644
index 00000000..83791ab8
--- /dev/null
+++ b/extra/elasticsearch/container_schema.json
@@ -0,0 +1,74 @@
+{
+"settings": {
+    "index": {
+        "analysis": {
+            "analyzer": {
+                "default": {
+                    "type": "custom",
+                    "tokenizer": "standard",
+                    "filter": [ "lowercase", "asciifolding" ]
+                },
+                "textIcu": {
+                    "type": "custom",
+                    "tokenizer": "icu_tokenizer",
+                    "char_filter": [ "icu_normalizer" ],
+                    "filter": [ "icu_folding" ]
+                },
+                "textIcuSearch": {
+                    "type": "custom",
+                    "tokenizer": "icu_tokenizer",
+                    "char_filter": [ "icu_normalizer" ],
+                    "filter": [ "icu_folding" ]
+                }
+            }
+        }
+    }
+},
+"mappings": {
+    "container": {
+        "properties": {
+            "ident":          { "type": "keyword" },
+            "state":          { "type": "keyword" },
+            "revision":       { "type": "keyword" },
+            "name":           { "type": "text", "index": true, "analyzer": "textIcu", "search_analyzer":"textIcuSearch" },
+            "publisher":      { "type": "text", "index": true, "analyzer": "textIcu", "search_analyzer":"textIcuSearch" },
+            "container_type": { "type": "keyword" },
+            "wikidata_qid":   { "type": "keyword" },
+            "issnl":          { "type": "keyword" },
+            "region":         { "type": "keyword" },
+            "nation":         { "type": "keyword" },
+            "discipline":     { "type": "keyword" },
+            "languages":      { "type": "keyword" },
+            "mimetypes":      { "type": "keyword" },
+            "first_year":     { "type": "integer" },
+            "last_year":      { "type": "integer" },
+
+            "in_doaj":        { "type": "boolean" },
+            "in_road":        { "type": "boolean" },
+            "in_doi":         { "type": "boolean" },
+            "in_doaj_works":  { "type": "boolean" },
+            "in_sherpa_romeo":{ "type": "boolean" },
+            "is_oa":          { "type": "boolean" },
+            "is_longtail_oa": { "type": "boolean" },
+            "any_kbart":      { "type": "boolean" },
+            "any_jstor":      { "type": "boolean" },
+            "any_sim":        { "type": "boolean" },
+            "ia_homepage_status":       { "type": "boolean" },
+
+            "releases_total": { "type": "integer" },
+            "releases_kbart": { "type": "integer" },
+            "releases_ia":    { "type": "integer" },
+            "releases_sim":   { "type": "integer" },
+            "releases_shadow":          { "type": "integer" },
+            "releases_any_file":        { "type": "integer" },
+            "releases_any_fileset":     { "type": "integer" },
+            "releases_any_webcapture":  { "type": "integer" },
+
+            "year":           { "type": "alias", "path": "first_year" },
+            "type":           { "type": "alias", "path": "container_type" },
+            "oa":             { "type": "alias", "path": "is_oa" },
+            "longtail":       { "type": "alias", "path": "is_longtail_oa" }
+        }
+    }
+}
+}