TS Lexi Tools

LexiTR API Documentation

API endpoints for frequency, collocation, and reverse dictionary queries.

1. Word Frequency API

Returns frequency data for a given word. Queries can target the exact word form or words that start with the given key.

GET /freq/frequency

Parameter	Required	Values	Description
`word`	Yes	Any Turkish word	The word or prefix to search.
`query_type`	Yes	`as_is`, `starts_with`	Controls exact-form or prefix frequency lookup.

$ curl "http://lexitr.tscorpus.com/freq/frequency?word=akıl&query_type=as_is"

# Example response
{
  "word": "akıl",
  "query_type": "as_is",
  "genre_frequencies": [
    {
      "word": "akıl",
      "academic_freq": 120,
      "social_media_freq": 840,
      "informative_freq": 530,
      "fictional_freq": 410,
      "total_freq": 1900
    }
  ]
}

$ curl "http://lexitr.tscorpus.com/freq/frequency?word=ak&query_type=starts_with"

2. Collocation API

Collocation endpoints return case-insensitive node-word matches, positional collocation scores, genre profiles, and corpus attestations for selected collocates.

2.1 Collocation Scores

GET POST JSON /col/collocations

Parameter	Required	Values	Description
`word`	Yes	Any Turkish word	The node word for collocation lookup. Matching is Turkish case-insensitive.
`genre`	No	`all`, `academic`, `informative`, `fictional`, `social_media`	Corpus genre filter. Defaults to `all`.
`span`	Yes	`1`-`5`	Number of tokens included on the left and right side of the node word.
`type`	No	`raw`, `mi`, `association`, `direction`, `all`	Controls the primary score block. Defaults to `all`.
`filter`	No	`Standart_SW`, `Greedy_SW`, `standard`, `greedy`, `not_use`, `none`	Stop-word filter applied to collocation candidates. Defaults to `Greedy_SW`.

$ curl "http://lexitr.tscorpus.com/col/collocations?word=akıl&genre=all&span=2&type=all&filter=Greedy_SW"

# Example response
{
  "word": "akıl",
  "genre": "all",
  "span": 2,
  "type": "all",
  "filter": "Greedy_SW",
  "normalized_filter": "greedy",
  "matching_lines": 1250,
  "matching_sentences": 1250,
  "node_occurrences": 1268,
  "left_context_tokens": 2500,
  "right_context_tokens": 2500,
  "corpus_tokens": 195603172,
  "score_method": "positional_log_ratio_v2",
  "direction_bias_method": "(right-left)/(right+left)",
  "raw_scores": [
    {
      "word": "baş",
      "left_frequency": 12,
      "right_frequency": 8,
      "overall_frequency": 4200
    }
  ],
  "mi_scores": [
    {
      "word": "baş",
      "left_frequency": 12,
      "left_mi": 1.42,
      "right_frequency": 8,
      "right_mi": 0.97
    }
  ],
  "association_scores": [
    {
      "word": "baş",
      "left_frequency": 12,
      "left_score": 1.31,
      "right_frequency": 8,
      "right_score": 0.86
    }
  ],
  "direction_bias_scores": [
    {
      "word": "baş",
      "left_frequency": 12,
      "right_frequency": 8,
      "direction_bias": -0.2,
      "direction": "left",
      "left_share": 0.6,
      "right_share": 0.4
    }
  ],
  "candidate_metrics": []
}

$ curl "http://lexitr.tscorpus.com/col/collocations?word=akıl&genre=academic&span=2&type=raw&filter=Standart_SW"

$ curl "http://lexitr.tscorpus.com/col/collocations?word=akıl&genre=fictional&span=2&type=mi&filter=none"

$ curl -X POST "http://lexitr.tscorpus.com/col/collocations" \
  -H "Content-Type: application/json" \
  -d '{"word":"akıl","genre":"all","span":2,"type":"all","filter":"Greedy_SW"}'

2.2 Collocate Genre Profile

Returns genre-level counts and normalized rates for a selected query-word / collocate pair.

GET POST JSON /col/genre-profile

Parameter	Required	Values	Description
`word`	Yes	Any Turkish word	Query node word. `query_word` is also accepted.
`collocate`	Yes	Any returned collocate	Collocate to profile. `collocate_word` is also accepted.
`span`	Yes	`1`-`5`	Position window used to count left and right occurrences.

$ curl "http://lexitr.tscorpus.com/col/genre-profile?word=akıl&collocate=baş&span=2"

# Example response
{
  "query_word": "akıl",
  "collocate_word": "baş",
  "span": 2,
  "genres": [
    {
      "genre": "informative",
      "corpus_tokens": 91785363,
      "query_frequency": 530,
      "collocate_frequency": 4200,
      "left_frequency": 12,
      "right_frequency": 8,
      "total_frequency": 20,
      "left_per_million": 0.13,
      "right_per_million": 0.09,
      "total_per_million": 0.22,
      "pair_per_1000_query_tokens": 37.74,
      "share_pct": 42.55,
      "direction_bias": -0.2,
      "matching_sentences": 19
    }
  ],
  "totals": {
    "left_frequency": 25,
    "right_frequency": 22,
    "total_frequency": 47,
    "direction_bias": -0.0638
  }
}

2.3 Positional Corpus Attestations

Returns paginated example sentences for a specific positional heatmap cell, such as a collocate at L1 or R2 of the query word.

GET /col/positional-attestations

Parameter	Required	Values	Description
`word`	Yes	Any Turkish word	Query node word.
`collocate`	Yes	Any returned collocate	Collocate to locate near the node word.
`side`	Yes	`left`, `right`	Whether the collocate must appear before or after the node.
`distance`	Yes	`1`-`5`	Exact token distance from the node word.
`genre`	No	`all`, `academic`, `informative`, `fictional`, `social_media`	Corpus genre filter. Defaults to `all`.
`page`	No	Positive integer	Result page. Defaults to `1`.
`page_size`	No	`1`-`100`	Attestations per page. Defaults to `20`.
`expected_count`	No	Integer	Optional heatmap count used to verify the retrieved attestation count.

$ curl "http://lexitr.tscorpus.com/col/positional-attestations?word=akıl&collocate=baş&side=left&distance=1&page=1&page_size=5"

# Example response
{
  "query_word": "akıl",
  "collocate_word": "baş",
  "side": "left",
  "distance": 1,
  "position": "L1",
  "genre": "all",
  "page": 1,
  "page_size": 5,
  "pages": 3,
  "total_attestations": 12,
  "total_sentences": 12,
  "retrieval_mode": "pair_fulltext",
  "heatmap_expected_count": null,
  "count_consistent": true,
  "examples": [
    {
      "attestation_number": 1,
      "sentence_id": 12345,
      "sentence": "...",
      "tokens": ["..."],
      "genre": "informative",
      "node_index": 8,
      "collocate_index": 7
    }
  ]
}

2.4 Single Word Frequency Helper

Returns the same genre-frequency distribution used by the Collocationary interface for an inline word-frequency chart.

POST JSON /col/trigger-function

Parameter	Required	Values	Description
`word`	Yes	Any Turkish word	Word form to look up in the frequency table.

$ curl -X POST "http://lexitr.tscorpus.com/col/trigger-function" \
  -H "Content-Type: application/json" \
  -d '{"word":"akıl"}'

3. Reverse Dictionary API

Returns words that end with a given key. For example, key=mut can return words such as armut and bizmut.

GET POST JSON /reverse/search

Parameter	Required	Values	Description
`key`	Yes	Ending sequence	The final letters or syllable sequence to match. `syl` is also accepted as an alias.
`w_type`	No	`All`, `Single_Word`, `Multi_Word`	Word type filter. Defaults to `All`.
`source`	No	`All`, `TS_Corpus`, `TDK_Dictionary`	Source filter. Defaults to `All`.
`limit`	No	1-1000	Maximum number of returned matches. Defaults to `500`.

$ curl "http://lexitr.tscorpus.com/reverse/search?key=mut&w_type=All&source=All&limit=50"

# Example response
{
  "key": "mut",
  "w_type": "All",
  "source": "All",
  "limit": 50,
  "match_count": 2,
  "results": [
    {
      "id": 101,
      "word": "armut",
      "w_type": "Single_Word",
      "s_count": 2,
      "tagged": "ar+VC mut+CVC",
      "s_type": "open",
      "s_kind": "N",
      "source": "TDK_Dictionary"
    }
  ]
}

$ curl -X POST "http://lexitr.tscorpus.com/reverse/search" \
  -H "Content-Type: application/json" \
  -d '{"key":"mut","w_type":"Single_Word","source":"TDK_Dictionary","limit":25}'