a3x/nyaa
1
0
Fork 0
mirror of https://gitlab.com/SIGBUS/nyaa.git synced 2024-11-05 13:55:54 +00:00
Code Issues Projects Releases Wiki Activity
nyaa/es_mapping.yml

140 lines
3.6 KiB
YAML
Raw Normal View History

some more elasticsearch work, including index mapping and analyzer
2017-05-15 18:14:01 +00:00
---
# CREATE DTABASE/TABLE equivalent for elasticsearch, in yaml
# fo inline comments.
settings:
analysis:
analyzer:
my_search_analyzer:
type: custom
tokenizer: standard
char_filter:
- my_char_filter
filter:
- standard
- lowercase
my_index_analyzer:
type: custom
tokenizer: standard
char_filter:
- my_char_filter
filter:
updated indicies
2017-05-18 08:58:08 +00:00
- resolution
some more elasticsearch work, including index mapping and analyzer
2017-05-15 18:14:01 +00:00
- lowercase
Resolves #129 and refactored create magnet es naming
2017-05-25 06:19:08 +00:00
- word_delimit
ES: delimit words before ngram, optimize tokens (#487) Before, long.tokens.with.dots.or.dashes would get edgengrammed up to the ngram limit, so we'd get to long.tokens.wit which would then be split - discarding "with.dots.or.dashes" completely. The fullword index would keep the complete large token, but without any ngramming, so incomplete searches (like "tokens") would not match it, only the full token. Now, we split words before ngramming them, so the main index will properly handle words up to the ngram limit. The fullword index will still handle the longer words for non-ngram matching. Also optimized away duplicate tokens from the indices (since we rely on boolean matching, not scoring) to save a couple megabytes of space.
2018-04-29 01:09:40 +00:00
- my_ngram
Fixes #227
2017-06-05 06:03:32 +00:00
- trim_zero
ES: delimit words before ngram, optimize tokens (#487) Before, long.tokens.with.dots.or.dashes would get edgengrammed up to the ngram limit, so we'd get to long.tokens.wit which would then be split - discarding "with.dots.or.dashes" completely. The fullword index would keep the complete large token, but without any ngramming, so incomplete searches (like "tokens") would not match it, only the full token. Now, we split words before ngramming them, so the main index will properly handle words up to the ngram limit. The fullword index will still handle the longer words for non-ngram matching. Also optimized away duplicate tokens from the indices (since we rely on boolean matching, not scoring) to save a couple megabytes of space.
2018-04-29 01:09:40 +00:00
- unique
[ES Change] Improve Elasticsearch term quoting (#473) * Optimize Elasticsearch fullword field Since the main display_name field ngrams words up to 15 characters, anything to and under that will already be indexed - the fullword field (which we have for words longer than 15 characters) needs to index only words longer than that. * Preprocess ES terms for better literal matching This commit adds a new .exact subfield to display_name, which holds a barely-filtered version of the original title we can do "literal" matching against. This is not real substring matching, but quoting terms now actually does something! Implements a simple preprocessor for the search terms to extract quoted parts from the search terms, optionally prefixed with - to negate them. The preprocessor will create a query that'll join all three query-types: the simple_query_string, must-phrases and must-not-phrases.
2018-04-14 00:06:25 +00:00
# For exact matching - simple lowercase + whitespace delimiter
exact_analyzer:
tokenizer: whitespace
filter:
- lowercase
[ES Schema change] Multi-field search display_name to match words over ngram limit This fixes searching for "Machiavellianism", 16 chars ("Machiavellianis", 15 chars, worked previously). Does not (seem to!) break anything, but requires a re-indexing of ES.
2017-06-05 14:29:00 +00:00
# For matching full words longer than the ngram limit (15 chars)
my_fullword_index_analyzer:
type: custom
tokenizer: standard
char_filter:
- my_char_filter
filter:
- lowercase
- word_delimit
[ES Change] Improve Elasticsearch term quoting (#473) * Optimize Elasticsearch fullword field Since the main display_name field ngrams words up to 15 characters, anything to and under that will already be indexed - the fullword field (which we have for words longer than 15 characters) needs to index only words longer than that. * Preprocess ES terms for better literal matching This commit adds a new .exact subfield to display_name, which holds a barely-filtered version of the original title we can do "literal" matching against. This is not real substring matching, but quoting terms now actually does something! Implements a simple preprocessor for the search terms to extract quoted parts from the search terms, optionally prefixed with - to negate them. The preprocessor will create a query that'll join all three query-types: the simple_query_string, must-phrases and must-not-phrases.
2018-04-14 00:06:25 +00:00
# Skip tokens shorter than N characters,
# since they're already indexed in the main field
- fullword_min
ES: delimit words before ngram, optimize tokens (#487) Before, long.tokens.with.dots.or.dashes would get edgengrammed up to the ngram limit, so we'd get to long.tokens.wit which would then be split - discarding "with.dots.or.dashes" completely. The fullword index would keep the complete large token, but without any ngramming, so incomplete searches (like "tokens") would not match it, only the full token. Now, we split words before ngramming them, so the main index will properly handle words up to the ngram limit. The fullword index will still handle the longer words for non-ngram matching. Also optimized away duplicate tokens from the indices (since we rely on boolean matching, not scoring) to save a couple megabytes of space.
2018-04-29 01:09:40 +00:00
- unique
[ES Schema change] Multi-field search display_name to match words over ngram limit This fixes searching for "Machiavellianism", 16 chars ("Machiavellianis", 15 chars, worked previously). Does not (seem to!) break anything, but requires a re-indexing of ES.
2017-06-05 14:29:00 +00:00
some more elasticsearch work, including index mapping and analyzer
2017-05-15 18:14:01 +00:00
filter:
my_ngram:
type: edgeNGram
min_gram: 1
max_gram: 15
[ES Change] Improve Elasticsearch term quoting (#473) * Optimize Elasticsearch fullword field Since the main display_name field ngrams words up to 15 characters, anything to and under that will already be indexed - the fullword field (which we have for words longer than 15 characters) needs to index only words longer than that. * Preprocess ES terms for better literal matching This commit adds a new .exact subfield to display_name, which holds a barely-filtered version of the original title we can do "literal" matching against. This is not real substring matching, but quoting terms now actually does something! Implements a simple preprocessor for the search terms to extract quoted parts from the search terms, optionally prefixed with - to negate them. The preprocessor will create a query that'll join all three query-types: the simple_query_string, must-phrases and must-not-phrases.
2018-04-14 00:06:25 +00:00
fullword_min:
type: length
# Remember to change this if you change the max_gram below!
min: 16
updated indicies
2017-05-18 08:58:08 +00:00
resolution:
type: pattern_capture
Resolves #129 and refactored create magnet es naming
2017-05-25 06:19:08 +00:00
patterns: ["(\\d+)[xX](\\d+)"]
Fixes #227
2017-06-05 06:03:32 +00:00
trim_zero:
type: pattern_capture
patterns: ["0*([0-9]*)"]
Resolves #129 and refactored create magnet es naming
2017-05-25 06:19:08 +00:00
word_delimit:
type: word_delimiter
preserve_original: true
split_on_numerics: false
some more elasticsearch work, including index mapping and analyzer
2017-05-15 18:14:01 +00:00
char_filter:
my_char_filter:
type: mapping
updated indicies
2017-05-18 08:58:08 +00:00
mappings: ["-=>_", "!=>_", "_=>\\u0020"]
some more elasticsearch work, including index mapping and analyzer
2017-05-15 18:14:01 +00:00
index:
# we're running a single es node, so no sharding necessary,
# plus replicas don't really help either.
number_of_shards: 1
number_of_replicas : 0
mapper:
# disable elasticsearch's "helpful" autoschema
dynamic: false
# since we disabled the _all field, default query the
# name of the torrent.
query:
default_field: display_name
mappings:
torrent:
# don't want everything concatenated
_all:
enabled: false
properties:
id:
type: long
display_name:
# TODO could do a fancier tokenizer here to parse out the
# the scene convention of stuff in brackets, plus stuff like k-on
type: text
analyzer: my_index_analyzer
[ES Schema change] Multi-field search display_name to match words over ngram limit This fixes searching for "Machiavellianism", 16 chars ("Machiavellianis", 15 chars, worked previously). Does not (seem to!) break anything, but requires a re-indexing of ES.
2017-06-05 14:29:00 +00:00
fielddata: true # Is this required?
fields:
# Multi-field for full-word matching (when going over ngram limits)
# Note: will have to be queried for, not automatic
fullword:
type: text
analyzer: my_fullword_index_analyzer
[ES Change] Improve Elasticsearch term quoting (#473) * Optimize Elasticsearch fullword field Since the main display_name field ngrams words up to 15 characters, anything to and under that will already be indexed - the fullword field (which we have for words longer than 15 characters) needs to index only words longer than that. * Preprocess ES terms for better literal matching This commit adds a new .exact subfield to display_name, which holds a barely-filtered version of the original title we can do "literal" matching against. This is not real substring matching, but quoting terms now actually does something! Implements a simple preprocessor for the search terms to extract quoted parts from the search terms, optionally prefixed with - to negate them. The preprocessor will create a query that'll join all three query-types: the simple_query_string, must-phrases and must-not-phrases.
2018-04-14 00:06:25 +00:00
# Stored for exact phrase matching
exact:
type: text
analyzer: exact_analyzer
some more elasticsearch work, including index mapping and analyzer
2017-05-15 18:14:01 +00:00
created_time:
type: date
# Only in the ES index for generating magnet links
info_hash:
enabled: false
filesize:
type: long
anonymous:
type: boolean
trusted:
type: boolean
remake:
type: boolean
complete:
type: boolean
hidden:
type: boolean
deleted:
type: boolean
has_torrent:
type: boolean
download_count:
type: long
leech_count:
type: long
seed_count:
type: long
Update ElasticSeach index and scripts for comment_count
2017-05-26 13:12:47 +00:00
comment_count:
type: long
some more elasticsearch work, including index mapping and analyzer
2017-05-15 18:14:01 +00:00
# these ids are really only for filtering, thus keyword
uploader_id:
type: keyword
main_category_id:
type: keyword
sub_category_id:
Reverted previous commit for mapping
2017-05-17 05:53:03 +00:00
type: keyword
Reference in a new issue Copy permalink