Update app.py
Browse files
app.py
CHANGED
|
@@ -24,19 +24,22 @@ def timer(func):
|
|
| 24 |
return detected_lang, confidence, execution_times
|
| 25 |
return detection
|
| 26 |
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
|
|
|
|
|
|
|
|
|
| 40 |
|
| 41 |
@timer
|
| 42 |
def detect_language(input_text: str, used_libraries: list[str]) -> tuple[str, str]:
|
|
@@ -56,13 +59,14 @@ def detect_language(input_text: str, used_libraries: list[str]) -> tuple[str, st
|
|
| 56 |
("en", 1.0)
|
| 57 |
"""
|
| 58 |
print(used_libraries)
|
|
|
|
| 59 |
detections = []
|
| 60 |
if 'langdetect' in used_libraries:
|
| 61 |
-
langcode, confidence_score = langdetect(input_text)
|
| 62 |
tupletoappend = langcode, confidence_score
|
| 63 |
detections.append(tupletoappend)
|
| 64 |
if 'langid' in used_libraries:
|
| 65 |
-
langcode, confidence_score = langid(input_text)
|
| 66 |
tupletoappend = langcode, confidence_score
|
| 67 |
detections.append(tupletoappend)
|
| 68 |
return detections
|
|
|
|
| 24 |
return detected_lang, confidence, execution_times
|
| 25 |
return detection
|
| 26 |
|
| 27 |
+
class Detect():
|
| 28 |
+
def __init__(self, text: str) -> None:
|
| 29 |
+
self.text: str = text
|
| 30 |
+
def langdetect(self) -> tuple[str, float]:
|
| 31 |
+
from langdetect import detect, detect_langs
|
| 32 |
+
from langdetect import DetectorFactory
|
| 33 |
+
DetectorFactory.seed = 0
|
| 34 |
+
langcode = detect(self.text)
|
| 35 |
+
langecode_probabilities: list[Language] = detect_langs(self.text)
|
| 36 |
+
return langcode, round(number=langecode_probabilities[0].prob * 100, ndigits=2)
|
| 37 |
+
|
| 38 |
+
def langid(self) -> tuple[str, float]:
|
| 39 |
+
import langid
|
| 40 |
+
result: tuple[str, float] = langid.classify(self.text)
|
| 41 |
+
langcode, langecode_probabilities = result
|
| 42 |
+
return langcode, round(number=langecode_probabilities * 10, ndigits=2)
|
| 43 |
|
| 44 |
@timer
|
| 45 |
def detect_language(input_text: str, used_libraries: list[str]) -> tuple[str, str]:
|
|
|
|
| 59 |
("en", 1.0)
|
| 60 |
"""
|
| 61 |
print(used_libraries)
|
| 62 |
+
detectinstance = Detect(input_text)
|
| 63 |
detections = []
|
| 64 |
if 'langdetect' in used_libraries:
|
| 65 |
+
langcode, confidence_score = detectinstance.langdetect(input_text)
|
| 66 |
tupletoappend = langcode, confidence_score
|
| 67 |
detections.append(tupletoappend)
|
| 68 |
if 'langid' in used_libraries:
|
| 69 |
+
langcode, confidence_score = detectinstance.langid(input_text)
|
| 70 |
tupletoappend = langcode, confidence_score
|
| 71 |
detections.append(tupletoappend)
|
| 72 |
return detections
|