Spaces:
Runtime error
Runtime error
removed redundant code
Browse files
summarizer/coreference_handler.py
DELETED
|
@@ -1,36 +0,0 @@
|
|
| 1 |
-
# removed previous import and related functionality since it's just a blank language model,
|
| 2 |
-
# while neuralcoref requires passing pretrained language model via spacy.load()
|
| 3 |
-
|
| 4 |
-
import neuralcoref
|
| 5 |
-
import spacy
|
| 6 |
-
|
| 7 |
-
from summarizer.sentence_handler import SentenceHandler
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
class CoreferenceHandler(SentenceHandler):
|
| 11 |
-
|
| 12 |
-
def __init__(self, spacy_model: str = 'en_core_web_sm',
|
| 13 |
-
greedyness: float = 0.45):
|
| 14 |
-
"""
|
| 15 |
-
Corefence handler. Only works with spacy < 3.0.
|
| 16 |
-
|
| 17 |
-
:param spacy_model: The spacy model to use as default.
|
| 18 |
-
:param greedyness: The greedyness factor.
|
| 19 |
-
"""
|
| 20 |
-
self.nlp = spacy.load(spacy_model)
|
| 21 |
-
neuralcoref.add_to_pipe(self.nlp, greedyness=greedyness)
|
| 22 |
-
|
| 23 |
-
def process(self, body: str, min_length: int = 40, max_length: int = 600):
|
| 24 |
-
"""
|
| 25 |
-
Processes the content sentences.
|
| 26 |
-
|
| 27 |
-
:param body: The raw string body to process
|
| 28 |
-
:param min_length: Minimum length that the sentences must be
|
| 29 |
-
:param max_length: Max length that the sentences mus fall under
|
| 30 |
-
:return: Returns a list of sentences.
|
| 31 |
-
"""
|
| 32 |
-
doc = self.nlp(body)._.coref_resolved
|
| 33 |
-
doc = self.nlp(doc)
|
| 34 |
-
return [c.string.strip()
|
| 35 |
-
for c in doc.sents
|
| 36 |
-
if max_length > len(c.string.strip()) > min_length]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|