From f6f7450903bdbe36bd5fff146b942e34ad221557 Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Fri, 25 Jan 2019 18:41:33 -0800 Subject: transform and import fixes/tweaks --- extra/elasticsearch/README.md | 16 +++-- extra/elasticsearch/changelog_schema.json | 5 +- extra/elasticsearch/fatcat_schema.json | 109 ++++++++++++++++++++++++++++++ 3 files changed, 122 insertions(+), 8 deletions(-) create mode 100644 extra/elasticsearch/fatcat_schema.json (limited to 'extra/elasticsearch') diff --git a/extra/elasticsearch/README.md b/extra/elasticsearch/README.md index 761ad6ab..691c6ed5 100644 --- a/extra/elasticsearch/README.md +++ b/extra/elasticsearch/README.md @@ -38,13 +38,17 @@ There is a Dockerfile in this directory which includes this installation. Drop and rebuild the schema: - http delete :9200/fatcat - http put :9200/fatcat < release_schema.json + http delete :9200/fatcat_release + http delete :9200/fatcat_container + http delete :9200/fatcat_changelog + http put :9200/fatcat_release < release_schema.json + http put :9200/fatcat_container < container_schema.json + http put :9200/fatcat_changelog < changelog_schema.json Put a single object (good for debugging): - head -n1 examples.json | http post :9200/fatcat/release/0 - http get :9200/fatcat/release/0 + head -n1 examples.json | http post :9200/fatcat_release/release/0 + http get :9200/fatcat_release/release/0 Bulk insert from a file on disk: @@ -53,14 +57,14 @@ Bulk insert from a file on disk: Or, in a bulk production live-stream conversion: export LC_ALL=C.UTF-8 - time zcat /srv/fatcat/snapshots/fatcat_release_dump_expanded.json.gz | ./fatcat_export.py transform-releases - - | esbulk -verbose -size 20000 -id ident -w 8 -index fatcat -type release + time zcat /srv/fatcat_release/snapshots/fatcat_release_dump_expanded.json.gz | ./fatcat_export.py transform-releases - - | esbulk -verbose -size 20000 -id ident -w 8 -index fatcat -type release ## Full-Text Querying A generic full-text "query string" query look like this (replace "blood" with actual query string, and "size" field with the max results to return): - GET /fatcat/release/_search + GET /fatcat_release/release/_search { "query": { "query_string": { diff --git a/extra/elasticsearch/changelog_schema.json b/extra/elasticsearch/changelog_schema.json index 7a7ec90c..f3211e99 100644 --- a/extra/elasticsearch/changelog_schema.json +++ b/extra/elasticsearch/changelog_schema.json @@ -16,8 +16,9 @@ "changelog": { "properties": { "index": { "type": "integer" }, - "editgorup_id": { "type": "keyword" }, - "timestamp": { "type": "datetime" }, + "editgroup_id": { "type": "keyword" }, + "timestamp": { "type": "date" }, + "editor_id": { "type": "keyword" }, "username": { "type": "keyword" }, "is_bot": { "type": "boolean" }, "is_admin": { "type": "boolean" }, diff --git a/extra/elasticsearch/fatcat_schema.json b/extra/elasticsearch/fatcat_schema.json new file mode 100644 index 00000000..05583330 --- /dev/null +++ b/extra/elasticsearch/fatcat_schema.json @@ -0,0 +1,109 @@ +{ +"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": { + "release": { + "properties": { + "ident": { "type": "keyword" }, + "state": { "type": "keyword" }, + "revision": { "type": "keyword" }, + "title": { "type": "text", "index": true, "analyzer": "textIcu", "search_analyzer":"textIcuSearch" }, + "original_title": { "type": "text", "index": true, "analyzer": "textIcu", "search_analyzer":"textIcuSearch" }, + "release_date": { "type": "date" }, + "release_year": { "type": "integer" }, + "release_type": { "type": "keyword" }, + "release_status": { "type": "keyword" }, + "language": { "type": "keyword" }, + "doi": { "type": "keyword" }, + "pmid": { "type": "keyword" }, + "pmcid": { "type": "keyword" }, + "isbn13": { "type": "keyword" }, + "wikidata_qid": { "type": "keyword" }, + "core_id": { "type": "keyword" }, + "axiv_id": { "type": "keyword" }, + "jstor_id": { "type": "keyword" }, + "license": { "type": "keyword" }, + "publisher": { "type": "text", "index": true, "analyzer": "textIcu", "search_analyzer":"textIcuSearch" }, + "container_name": { "type": "text", "index": true, "analyzer": "textIcu", "search_analyzer":"textIcuSearch" }, + "container_issnl": { "type": "keyword" }, + "container_type": { "type": "keyword" }, + "contrib_count": { "type": "integer" }, + "contrib_names": { "type": "text", "index": true, "analyzer": "textIcu", "search_analyzer":"textIcuSearch" }, + "ref_count": { "type": "integer" }, + "file_count": { "type": "integer" }, + "fileset_count": { "type": "integer" }, + "webcapture_count": { "type": "integer" }, + "any_abstract": { "type": "boolean" }, + + "best_pdf_url": { "type": "keyword" }, + "ia_pdf_url": { "type": "keyword" }, + "is_oa": { "type": "boolean" }, + "is_longtail_oa": { "type": "boolean" }, + "is_preserved": { "type": "boolean" }, + "in_kbart": { "type": "boolean" }, + "in_jstor": { "type": "boolean" }, + "in_dweb": { "type": "boolean" }, + "in_web": { "type": "boolean" }, + "in_ia": { "type": "boolean" }, + "in_ia_sim": { "type": "boolean" }, + "in_shadows": { "type": "boolean" }, + + "author": { "type": "alias", "path": "contrib_names" }, + "journal": { "type": "alias", "path": "container_name" }, + "date": { "type": "alias", "path": "release_date" }, + "year": { "type": "alias", "path": "release_year" }, + "issn": { "type": "alias", "path": "container_issnl" }, + "oa": { "type": "alias", "path": "is_oa" }, + "longtail": { "type": "alias", "path": "is_longtail_oa" }, + "lang": { "type": "alias", "path": "language" }, + "file_pdf_url": { "type": "alias", "path": "best_pdf_url" }, + "is_kept": { "type": "alias", "path": "in_kbart" } + } + }, + "changelog": { + "properties": { + "index": { "type": "integer" }, + "editgorup_id": { "type": "keyword" }, + "timestamp": { "type": "date" }, + "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" } + } + } +} +} -- cgit v1.2.3