LoloSemper commited on
Commit
c163687
·
verified ·
1 Parent(s): 6094523

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +31 -41
app.py CHANGED
@@ -386,7 +386,7 @@ def extract_custom_sidecar(text: str) -> Optional[str]:
386
  def strip_custom_sidecar(text: str) -> str:
387
  return text.split('~')[0].rstrip() if '~' in text else text
388
 
389
- # ------------ Codificar / Decodificar léxico puro ------------
390
  def encode_simple(text: str, src_lang: str, target: str) -> str:
391
  if not text.strip(): return ""
392
  def repl_es(m):
@@ -499,46 +499,36 @@ def decode_simple(text: str, source: str, tgt_lang: str) -> str:
499
  lemma_tokens.append(w)
500
  pl_flags.append(had_pl)
501
 
502
- if verb_idx == -1:
503
- # Fallback zero copula
504
- verb_lemma = "ser" if tgt_lang == "Español" else "be"
505
- verb_tense = "Pres"
506
- verb_person = "3s"
507
- v_conj = _es_conj(verb_lemma, verb_tense, verb_person) if tgt_lang == "Español" else _en_conj(verb_lemma, verb_tense, verb_person)
508
- lemma_tokens.insert(1 if lemma_tokens else 0, v_conj)
509
- out_text = " ".join(lemma_tokens)
510
- else:
511
- # Conjuga
512
- conj_func = _es_conj if tgt_lang == "Español" else _en_conj
513
- v_conj = conj_func(verb_lemma, verb_tense, verb_person)
514
- if is_neg:
515
- neg_prefix = "no " if tgt_lang == "Español" else "not "
516
- v_conj = neg_prefix + v_conj
517
-
518
- # Reordena SVO
519
- post_v = lemma_tokens[verb_idx + 1:]
520
- pl_post = pl_flags[verb_idx + 1:]
521
- s_idx = next((j for j, w in enumerate(post_v) if w.lower() in pron_set), None)
522
- S = post_v[s_idx] if s_idx is not None else None
523
- if S:
524
- if pl_post[s_idx]:
525
- S = pluralize(S, tgt_lang)
526
- del post_v[s_idx]
527
- del pl_post[s_idx]
528
-
529
- O_ADV = []
530
- if post_v:
531
- O = pluralize(post_v[0], tgt_lang) if pl_post[0] else post_v[0]
532
- O_ADV.append(O)
533
- O_ADV.extend([pluralize(post_v[k], tgt_lang) if pl_post[k] else post_v[k] for k in range(1, len(post_v))])
534
-
535
- parts = [p for p in [S, v_conj] + O_ADV if p]
536
- out_text = " ".join(parts)
537
-
538
- # Wh en Q: si primer token es wh, muévelo al frente
539
- if has_q and lemma_tokens and lemma_tokens[0].lower() in {"como", "cómo", "what", "how"}:
540
- wh = lemma_tokens.pop(0)
541
- out_text = f"{wh} {out_text}"
542
 
543
  # Pregunta
544
  if has_q:
 
386
  def strip_custom_sidecar(text: str) -> str:
387
  return text.split('~')[0].rstrip() if '~' in text else text
388
 
389
+ # ------------ Codificar / Decodificar léxisco puro ------------
390
  def encode_simple(text: str, src_lang: str, target: str) -> str:
391
  if not text.strip(): return ""
392
  def repl_es(m):
 
499
  lemma_tokens.append(w)
500
  pl_flags.append(had_pl)
501
 
502
+ out_parts = []
503
+ greeting = None
504
+ wh = None
505
+ for idx, lem in enumerate(lemma_tokens):
506
+ if idx == verb_idx:
507
+ conj_func = _es_conj if tgt_lang == "Español" else _en_conj
508
+ v_conj = conj_func(verb_lemma, verb_tense, verb_person)
509
+ if is_neg:
510
+ neg_prefix = "no " if tgt_lang == "Español" else "not "
511
+ v_conj = neg_prefix + v_conj
512
+ out_parts.append(v_conj)
513
+ continue
514
+
515
+ w = pluralize(lem, tgt_lang) if pl_flags[idx] else lem
516
+ if w.lower() in {"hola", "hello", "hi", "hey"}:
517
+ greeting = w
518
+ elif w.lower() in {"como", "cómo", "what", "how"} and has_q:
519
+ wh = w
520
+ else:
521
+ out_parts.append(w)
522
+
523
+ # Reorden: Greeting + wh + S V O ADV
524
+ final_out = []
525
+ if greeting:
526
+ final_out.append(greeting)
527
+ if wh:
528
+ final_out.append(wh)
529
+ final_out += out_parts
530
+
531
+ out_text = " ".join(final_out)
 
 
 
 
 
 
 
 
 
 
532
 
533
  # Pregunta
534
  if has_q: