Logic package

Subpackages

• Logic.core package
  • Subpackages
    • Logic.core.classification package
      • Logic.core.classification.basic_classifier module
      • Logic.core.classification.data_loader module
      • Logic.core.classification.deep module
      • Logic.core.classification.knn module
      • Logic.core.classification.naive_bayes module
      • Logic.core.classification.svm module
    • Logic.core.clustering package
      • Logic.core.clustering.clustering_metrics module
      • Logic.core.clustering.clustering_utils module
      • Logic.core.clustering.dimension_reduction module
      • Logic.core.clustering.main module
    • Logic.core.indexer package
      • Logic.core.indexer.LSH module
      • Logic.core.indexer.document_lengths_index module
      • Logic.core.indexer.index module
      • Logic.core.indexer.index_reader module
      • Logic.core.indexer.indexes_enum module
      • Logic.core.indexer.metadata_index module
      • Logic.core.indexer.tiered_index module
    • Logic.core.link_analysis package
      • Logic.core.link_analysis.analyzer module
      • Logic.core.link_analysis.graph module
    • Logic.core.utility package
      • Logic.core.utility.crawler module
      • Logic.core.utility.evaluation module
      • Logic.core.utility.preprocess module
      • Logic.core.utility.scorer module
      • Logic.core.utility.snippet module
      • Logic.core.utility.spell_correction module
    • Logic.core.word_embedding package
      • Logic.core.word_embedding.fasttext_data_loader module
      • Logic.core.word_embedding.fasttext_model module
  • Logic.core.search module
    • SearchEngine
      • SearchEngine.aggregate_scores()
      • SearchEngine.find_scores_with_safe_ranking()
      • SearchEngine.find_scores_with_unigram_model()
      • SearchEngine.find_scores_with_unsafe_ranking()
      • SearchEngine.merge_scores()
      • SearchEngine.search()

Logic.utils module

correct_text(text: str, all_documents: List[str]) → str

Correct the give query text, if it is misspelled using Jacard similarity

Paramters

text: str

The query text

all_documentslist of str

The input documents.

Returns str

The corrected form of the given text

get_movie_by_id(id: str, movies_dataset: List[Dict[str, str]]) → Dict[str, str]

Get movie by its id

Parameters:

• id (str) – The id of the movie

• movies_dataset (List[Dict[str, str]]) – The dataset of movies

Returns:

The movie with the given id

Return type:

dict

search(query: str, max_result_count: int, method: str = 'ltn-lnn', weights: list = [0.3, 0.3, 0.4], should_print=False, preferred_genre: str = None, smoothing_method=None, alpha=0.5, lamda=0.5)

Finds relevant documents to query

Parameters:

• query – The query text

• max_result_count (Return top 'max_result_count' docs which have the highest scores.) – notice that if max_result_count = -1, then you have to return all docs

• method ('ltn.lnn' or 'ltc.lnc' or 'OkapiBM25')

• weights –

  The list, containing importance weights in the search result for each of these items:

  Indexes.STARS: weights[0], Indexes.GENRES: weights[1], Indexes.SUMMARIES: weights[2],

• preferred_genre – A list containing preference rates for each genre. If None, the preference rates are equal. (You can leave it None for now)

Returns:

• list

• Retrieved documents with snippet