def test_at_least_one_of_the_tokens(self):
test_doc_list = [
TextDocument("there is a city named New York", "doc1"),
TextDocument("In Britain there is a city named York", "doc2")
]
expected_list = ['doc1', 'doc2']
small_collection = DocumentCollection.from_document_list(test_doc_list)
searcher = SearchEngine(small_collection)
query = "New York Britain"
top_docs = searcher.ranked_documents(query)
result_list = []
for doc, sim in top_docs:
result_list.append(doc.id)
self.assertEqual(result_list, expected_list)
def setUp(self):
test_doc_list = [
TextDocument(text_and_id[0], text_and_id[1])
for text_and_id in [("the cat sat on a mat",
"doc1"), ("a rose is a rose", "doc2")]
]
self.small_collection = DocumentCollection.from_document_list(
test_doc_list)
def test_unknown_word_cosine(self):
""" Return 0 if cosine similarity is called for documents with only out-of-vocabulary words. """
# Document that only contains words that never occurred in the document collection.
query_doc = TextDocument(text="unknownwords", id=None)
# Some document from collection.
collection_doc = self.small_collection.docid_to_doc["doc1"]
# Similarity should be zero (instead of undefined).
self.assertEqual(
self.small_collection.cosine_similarity(query_doc, collection_doc),
0.)
def setUp(self):
test_doc_list = [TextDocument("the cat sat\non a mat", "doc1")]
self.small_collection = DocumentCollection.from_document_list(
test_doc_list)
self.searcher = SearchEngine(self.small_collection)
def setUp(self):
self.test_reply_email = TextDocument(">>>>forword", "doc1")