提交

kongdeqiang

2026-03-13 28ccfbfc51068a663a80764e14074df5edf2b5ba

 funasr/metrics/wer.py

@@ -5,186 +5,221 @@
from omegaconf import DictConfig, OmegaConf, ListConfig


def compute_wer(ref_file,
                hyp_file,
                cer_file,
                cn_postprocess=False,
                ):
   rst = {
      'Wrd': 0,
      'Corr': 0,
      'Ins': 0,
      'Del': 0,
      'Sub': 0,
      'Snt': 0,
      'Err': 0.0,
      'S.Err': 0.0,
      'wrong_words': 0,
      'wrong_sentences': 0
   }
	
   hyp_dict = {}
   ref_dict = {}
   with open(hyp_file, 'r') as hyp_reader:
      for line in hyp_reader:
         key = line.strip().split()[0]
         value = line.strip().split()[1:]
         if cn_postprocess:
            value = " ".join(value)
            value = value.replace(" ", "")
            if value[0] == "请":
               value = value[1:]
            value = [x for x in value]
         hyp_dict[key] = value
   with open(ref_file, 'r') as ref_reader:
      for line in ref_reader:
         key = line.strip().split()[0]
         value = line.strip().split()[1:]
         if cn_postprocess:
            value = " ".join(value)
            value = value.replace(" ", "")
            value = [x for x in value]
         ref_dict[key] = value
	
   cer_detail_writer = open(cer_file, 'w')
   for hyp_key in hyp_dict:
      if hyp_key in ref_dict:
         out_item = compute_wer_by_line(hyp_dict[hyp_key], ref_dict[hyp_key])
         rst['Wrd'] += out_item['nwords']
         rst['Corr'] += out_item['cor']
         rst['wrong_words'] += out_item['wrong']
         rst['Ins'] += out_item['ins']
         rst['Del'] += out_item['del']
         rst['Sub'] += out_item['sub']
         rst['Snt'] += 1
         if out_item['wrong'] > 0:
            rst['wrong_sentences'] += 1
         cer_detail_writer.write(hyp_key + print_cer_detail(out_item) + '\n')
         cer_detail_writer.write("ref:" + '\t' + " ".join(list(map(lambda x: x.lower(), ref_dict[hyp_key]))) + '\n')
         cer_detail_writer.write("hyp:" + '\t' + " ".join(list(map(lambda x: x.lower(), hyp_dict[hyp_key]))) + '\n')
         cer_detail_writer.flush()
	
   if rst['Wrd'] > 0:
      rst['Err'] = round(rst['wrong_words'] * 100 / rst['Wrd'], 2)
   if rst['Snt'] > 0:
      rst['S.Err'] = round(rst['wrong_sentences'] * 100 / rst['Snt'], 2)
	
   cer_detail_writer.write('\n')
   cer_detail_writer.write("%WER " + str(rst['Err']) + " [ " + str(rst['wrong_words']) + " / " + str(rst['Wrd']) +
                           ", " + str(rst['Ins']) + " ins, " + str(rst['Del']) + " del, " + str(
      rst['Sub']) + " sub ]" + '\n')
   cer_detail_writer.write(
      "%SER " + str(rst['S.Err']) + " [ " + str(rst['wrong_sentences']) + " / " + str(rst['Snt']) + " ]" + '\n')
   cer_detail_writer.write("Scored " + str(len(hyp_dict)) + " sentences, " + str(
      len(hyp_dict) - rst['Snt']) + " not present in hyp." + '\n')
	
   cer_detail_writer.close()
def compute_wer(
    ref_file,
    hyp_file,
    cer_file,
    cn_postprocess=False,
):
    rst = {
        "Wrd": 0,
        "Corr": 0,
        "Ins": 0,
        "Del": 0,
        "Sub": 0,
        "Snt": 0,
        "Err": 0.0,
        "S.Err": 0.0,
        "wrong_words": 0,
        "wrong_sentences": 0,
    }

    hyp_dict = {}
    ref_dict = {}
    with open(hyp_file, "r") as hyp_reader:
        for line in hyp_reader:
            key = line.strip().split()[0]
            value = line.strip().split()[1:]
            if cn_postprocess:
                value = " ".join(value)
                value = value.replace(" ", "")
                # if value[0] == "请":
                #     value = value[1:]
                value = [x for x in value]
            hyp_dict[key] = value
    with open(ref_file, "r") as ref_reader:
        for line in ref_reader:
            key = line.strip().split()[0]
            value = line.strip().split()[1:]
            if cn_postprocess:
                value = " ".join(value)
                value = value.replace(" ", "")
                value = [x for x in value]
            ref_dict[key] = value

    cer_detail_writer = open(cer_file, "w")
    for hyp_key in hyp_dict:
        if hyp_key in ref_dict:
            out_item = compute_wer_by_line(hyp_dict[hyp_key], ref_dict[hyp_key])
            rst["Wrd"] += out_item["nwords"]
            rst["Corr"] += out_item["cor"]
            rst["wrong_words"] += out_item["wrong"]
            rst["Ins"] += out_item["ins"]
            rst["Del"] += out_item["del"]
            rst["Sub"] += out_item["sub"]
            rst["Snt"] += 1
            if out_item["wrong"] > 0:
                rst["wrong_sentences"] += 1
            cer_detail_writer.write(hyp_key + print_cer_detail(out_item) + "\n")
            cer_detail_writer.write(
                "ref:" + "\t" + " ".join(list(map(lambda x: x.lower(), ref_dict[hyp_key]))) + "\n"
            )
            cer_detail_writer.write(
                "hyp:" + "\t" + " ".join(list(map(lambda x: x.lower(), hyp_dict[hyp_key]))) + "\n"
            )
            cer_detail_writer.flush()

    if rst["Wrd"] > 0:
        rst["Err"] = round(rst["wrong_words"] * 100 / rst["Wrd"], 2)
    if rst["Snt"] > 0:
        rst["S.Err"] = round(rst["wrong_sentences"] * 100 / rst["Snt"], 2)

    cer_detail_writer.write("\n")
    cer_detail_writer.write(
        "%WER "
        + str(rst["Err"])
        + " [ "
        + str(rst["wrong_words"])
        + " / "
        + str(rst["Wrd"])
        + ", "
        + str(rst["Ins"])
        + " ins, "
        + str(rst["Del"])
        + " del, "
        + str(rst["Sub"])
        + " sub ]"
        + "\n"
    )
    cer_detail_writer.write(
        "%SER "
        + str(rst["S.Err"])
        + " [ "
        + str(rst["wrong_sentences"])
        + " / "
        + str(rst["Snt"])
        + " ]"
        + "\n"
    )
    cer_detail_writer.write(
        "Scored "
        + str(len(hyp_dict))
        + " sentences, "
        + str(len(hyp_dict) - rst["Snt"])
        + " not present in hyp."
        + "\n"
    )

    cer_detail_writer.close()


def compute_wer_by_line(hyp,
                        ref):
   hyp = list(map(lambda x: x.lower(), hyp))
   ref = list(map(lambda x: x.lower(), ref))
	
   len_hyp = len(hyp)
   len_ref = len(ref)
	
   cost_matrix = np.zeros((len_hyp + 1, len_ref + 1), dtype=np.int16)
	
   ops_matrix = np.zeros((len_hyp + 1, len_ref + 1), dtype=np.int8)
	
   for i in range(len_hyp + 1):
      cost_matrix[i][0] = i
   for j in range(len_ref + 1):
      cost_matrix[0][j] = j
	
   for i in range(1, len_hyp + 1):
      for j in range(1, len_ref + 1):
         if hyp[i - 1] == ref[j - 1]:
            cost_matrix[i][j] = cost_matrix[i - 1][j - 1]
         else:
            substitution = cost_matrix[i - 1][j - 1] + 1
            insertion = cost_matrix[i - 1][j] + 1
            deletion = cost_matrix[i][j - 1] + 1
				
            compare_val = [substitution, insertion, deletion]
				
            min_val = min(compare_val)
            operation_idx = compare_val.index(min_val) + 1
            cost_matrix[i][j] = min_val
            ops_matrix[i][j] = operation_idx
	
   match_idx = []
   i = len_hyp
   j = len_ref
   rst = {
      'nwords': len_ref,
      'cor': 0,
      'wrong': 0,
      'ins': 0,
      'del': 0,
      'sub': 0
   }
   while i >= 0 or j >= 0:
      i_idx = max(0, i)
      j_idx = max(0, j)
		
      if ops_matrix[i_idx][j_idx] == 0:  # correct
         if i - 1 >= 0 and j - 1 >= 0:
            match_idx.append((j - 1, i - 1))
            rst['cor'] += 1
			
         i -= 1
         j -= 1
		
      elif ops_matrix[i_idx][j_idx] == 2:  # insert
         i -= 1
         rst['ins'] += 1
		
      elif ops_matrix[i_idx][j_idx] == 3:  # delete
         j -= 1
         rst['del'] += 1
		
      elif ops_matrix[i_idx][j_idx] == 1:  # substitute
         i -= 1
         j -= 1
         rst['sub'] += 1
		
      if i < 0 and j >= 0:
         rst['del'] += 1
      elif j < 0 and i >= 0:
         rst['ins'] += 1
	
   match_idx.reverse()
   wrong_cnt = cost_matrix[len_hyp][len_ref]
   rst['wrong'] = wrong_cnt
	
   return rst
def compute_wer_by_line(hyp, ref):
    hyp = list(map(lambda x: x.lower(), hyp))
    ref = list(map(lambda x: x.lower(), ref))

    len_hyp = len(hyp)
    len_ref = len(ref)

    cost_matrix = np.zeros((len_hyp + 1, len_ref + 1), dtype=np.int16)

    ops_matrix = np.zeros((len_hyp + 1, len_ref + 1), dtype=np.int8)

    for i in range(len_hyp + 1):
        cost_matrix[i][0] = i
    for j in range(len_ref + 1):
        cost_matrix[0][j] = j

    for i in range(1, len_hyp + 1):
        for j in range(1, len_ref + 1):
            if hyp[i - 1] == ref[j - 1]:
                cost_matrix[i][j] = cost_matrix[i - 1][j - 1]
            else:
                substitution = cost_matrix[i - 1][j - 1] + 1
                insertion = cost_matrix[i - 1][j] + 1
                deletion = cost_matrix[i][j - 1] + 1

                compare_val = [substitution, insertion, deletion]

                min_val = min(compare_val)
                operation_idx = compare_val.index(min_val) + 1
                cost_matrix[i][j] = min_val
                ops_matrix[i][j] = operation_idx

    match_idx = []
    i = len_hyp
    j = len_ref
    rst = {"nwords": len_ref, "cor": 0, "wrong": 0, "ins": 0, "del": 0, "sub": 0}
    while i >= 0 or j >= 0:
        i_idx = max(0, i)
        j_idx = max(0, j)

        if ops_matrix[i_idx][j_idx] == 0:  # correct
            if i - 1 >= 0 and j - 1 >= 0:
                match_idx.append((j - 1, i - 1))
                rst["cor"] += 1

            i -= 1
            j -= 1

        elif ops_matrix[i_idx][j_idx] == 2:  # insert
            i -= 1
            rst["ins"] += 1

        elif ops_matrix[i_idx][j_idx] == 3:  # delete
            j -= 1
            rst["del"] += 1

        elif ops_matrix[i_idx][j_idx] == 1:  # substitute
            i -= 1
            j -= 1
            rst["sub"] += 1

        if i < 0 and j >= 0:
            rst["del"] += 1
        elif j < 0 and i >= 0:
            rst["ins"] += 1

    match_idx.reverse()
    wrong_cnt = cost_matrix[len_hyp][len_ref]
    rst["wrong"] = wrong_cnt

    return rst


def print_cer_detail(rst):
   return ("(" + "nwords=" + str(rst['nwords']) + ",cor=" + str(rst['cor'])
           + ",ins=" + str(rst['ins']) + ",del=" + str(rst['del']) + ",sub="
           + str(rst['sub']) + ") corr:" + '{:.2%}'.format(rst['cor'] / rst['nwords'])
           + ",cer:" + '{:.2%}'.format(rst['wrong'] / rst['nwords']))
    return (
        "("
        + "nwords="
        + str(rst["nwords"])
        + ",cor="
        + str(rst["cor"])
        + ",ins="
        + str(rst["ins"])
        + ",del="
        + str(rst["del"])
        + ",sub="
        + str(rst["sub"])
        + ") corr:"
        + "{:.2%}".format(rst["cor"] / rst["nwords"])
        + ",cer:"
        + "{:.2%}".format(rst["wrong"] / rst["nwords"])
    )


@hydra.main(config_name=None, version_base=None)
def main_hydra(cfg: DictConfig):
   ref_file = cfg.get("ref_file", None)
   hyp_file = cfg.get("hyp_file", None)
   cer_file = cfg.get("cer_file", None)
   cn_postprocess = cfg.get("cn_postprocess", False)
   if ref_file is None or hyp_file is None or cer_file is None:
      print(
         "usage : python -m  funasr.metrics.wer ++ref_file=test.ref ++hyp_file=test.hyp ++cer_file=test.wer ++cn_postprocess=false")
      sys.exit(0)
	
   compute_wer(ref_file, hyp_file, cer_file, cn_postprocess)
    ref_file = cfg.get("ref_file", None)
    hyp_file = cfg.get("hyp_file", None)
    cer_file = cfg.get("cer_file", None)
    cn_postprocess = cfg.get("cn_postprocess", False)
    if ref_file is None or hyp_file is None or cer_file is None:
        print(
            "usage : python -m  funasr.metrics.wer ++ref_file=test.ref ++hyp_file=test.hyp ++cer_file=test.wer ++cn_postprocess=false"
        )
        sys.exit(0)

    compute_wer(ref_file, hyp_file, cer_file, cn_postprocess)


if __name__ == '__main__':
   main_hydra()



if __name__ == "__main__":
    main_hydra()