IterModel

class matchup.models.model.IterModel

Bases: matchup.models.model.Model

Describe one variation of Model classes : IterModel classes have some features for help his works Pointers and occurrences are implemented here.

Methods Summary

doc_repr(doc)	Process doc generating it representation
initialize(query, vocabulary)	Initialize query-based
initialize_occurrences(query, vocabulary)	Create another data structure _term_occurrences that represents the vocabulary with just query
initialize_pointers()	Initialize pointers to model algorithm
iter()	Define one iteration of this iter model algorithm
next_doc()	Return the lowest doc pointer by pointers
process_vocabulary_query_based(query, vocabulary)	Generate document scores based in query
query_repr(query, idf, tf)	Construct query representation
run(query, vocabulary)	Define the principal method of IR models.
stop()	Given a dictionary with all pointers by keyword and another dictionary with all occurrences by keyword,

Methods Documentation

doc_repr(doc: str) → DefaultDict[str, float]

Process doc generating it representation

Parameters:doc – Str represents the lowest document Returns:doc repr. Dictionary with all term scores by doc. That is the document vector.

initialize(query: List[matchup.presentation.text.Term], vocabulary: matchup.structure.vocabulary.Vocabulary)

Initialize query-based

initialize_occurrences(query: List[matchup.presentation.text.Term], vocabulary: matchup.structure.vocabulary.Vocabulary) → None

Create another data structure _term_occurrences that represents the vocabulary with just query keywords.

Parameters:

• query – original query
• vocabulary – original vocabulary

Returns:

None

initialize_pointers() → None

Initialize pointers to model algorithm

Returns:None

iter() → Tuple[str, DefaultDict[str, float]]

Define one iteration of this iter model algorithm

Returns:doc, doc_repr (keyword -> score)

next_doc() → str

Return the lowest doc pointer by pointers

Returns:lowest document

process_vocabulary_query_based(query: List[matchup.presentation.text.Term], vocabulary: matchup.structure.vocabulary.Vocabulary) → DefaultDict[str, List[matchup.structure.occurrence.Occurrence]]

Generate document scores based in query

Parameters:

• query – query representation
• vocabulary – vocabulary structure

Returns:

List of occurrences

classmethod query_repr(query: List[matchup.presentation.text.Term], idf, tf) → DefaultDict[str, float]

Construct query representation

Parameters:

• query – list of all terms
• idf – structure IDF
• tf – structure TF

Returns:

query representation

run(query: List[matchup.presentation.text.Term], vocabulary: matchup.structure.vocabulary.Vocabulary) → List[matchup.structure.solution.Result]

Define the principal method of IR models.

Parameters:

• query – List of all entry terms
• vocabulary – Vocabulary pre-processed

Returns:

stop() → bool

Given a dictionary with all pointers by keyword and another dictionary with all occurrences by keyword, this function calculate if the algorithm is over

Returns:boolean flag indicates if algorithm stops