TymaaHammouda commited on
Commit
f43b328
·
1 Parent(s): 34a52a1

Update alma output

Browse files
Files changed (1) hide show
  1. app.py +18 -8
app.py CHANGED
@@ -4,7 +4,7 @@ import os
4
  from pydantic import BaseModel
5
  from fastapi.responses import JSONResponse
6
 
7
- print("Version ---- 1")
8
  app = FastAPI()
9
 
10
  def download_file_from_hf(repo_id, filename):
@@ -27,7 +27,7 @@ download_file_from_hf("SinaLab/ArabGlossBERT","four_grams.pickle")
27
  download_file_from_hf("SinaLab/ArabGlossBERT","five_grams.pickle")
28
 
29
  from sinatools.morphology.morph_analyzer import analyze
30
-
31
 
32
  class ALMARequest(BaseModel):
33
  text: str
@@ -42,12 +42,22 @@ def predict(request: ALMARequest):
42
  task = request.task
43
  flag = request.flag
44
 
45
- alma_output = analyze(text, language, task, str(flag))
46
- content = {
47
- "resp": alma_output,
48
- "statusText": "OK",
49
- "statusCode": 0,
50
- }
 
 
 
 
 
 
 
 
 
 
51
 
52
  return JSONResponse(
53
  content=content,
 
4
  from pydantic import BaseModel
5
  from fastapi.responses import JSONResponse
6
 
7
+ print("Version ---- 2")
8
  app = FastAPI()
9
 
10
  def download_file_from_hf(repo_id, filename):
 
27
  download_file_from_hf("SinaLab/ArabGlossBERT","five_grams.pickle")
28
 
29
  from sinatools.morphology.morph_analyzer import analyze
30
+ from sinatools.utils.tokenizer import sentence_tokenizer
31
 
32
  class ALMARequest(BaseModel):
33
  text: str
 
42
  task = request.task
43
  flag = request.flag
44
 
45
+ sentences = sentence_tokenizer(text, dot=True, new_line=True, question_mark=True, exclamation_mark=True)
46
+
47
+ results_with_sentences = []
48
+
49
+ for i, sentence in enumerate(sentences):
50
+ sentence_id = i + 1
51
+
52
+ result = analyze(sentence, language, task, str(flag))
53
+
54
+ results_with_sentences.append({
55
+ "sentence_id": sentence_id,
56
+ "sentence": sentence,
57
+ "lemmatizer_results": result
58
+ })
59
+
60
+ content = {"resp": results_with_sentences, "statusText": "OK", "statusCode": 0}
61
 
62
  return JSONResponse(
63
  content=content,