Spaces:
Runtime error
Runtime error
Updating layout and introducing interoperability
Browse files
app.py
CHANGED
@@ -1,21 +1,26 @@
|
|
1 |
import csv
|
2 |
-
import statistics
|
3 |
import string
|
4 |
|
5 |
import gensim.downloader as api
|
6 |
-
import
|
7 |
import nltk
|
|
|
8 |
import pandas as pd
|
9 |
import readability
|
10 |
import seaborn as sns
|
11 |
import torch
|
12 |
from fuzzywuzzy import fuzz
|
13 |
from nltk.corpus import stopwords
|
|
|
14 |
from nltk.tokenize import word_tokenize
|
15 |
from sklearn.metrics.pairwise import cosine_similarity
|
16 |
from transformers import DistilBertTokenizer
|
17 |
from transformers import pipeline
|
18 |
|
|
|
|
|
|
|
|
|
19 |
nltk.download('cmudict')
|
20 |
|
21 |
nltk.download('stopwords')
|
@@ -29,85 +34,102 @@ device = torch.device('cuda' if torch.cuda.is_available else 'cpu')
|
|
29 |
|
30 |
# loading model
|
31 |
PATH = 'pytorchBERTmodel'
|
32 |
-
model = torch.load(PATH
|
33 |
model.eval()
|
34 |
|
35 |
-
model.to(
|
36 |
|
37 |
p = pipeline("automatic-speech-recognition")
|
38 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
w2v = dict({})
|
40 |
-
for idx, key in enumerate(glove_vectors.
|
41 |
-
w2v[key] = glove_vectors.get_vector(key)
|
42 |
|
43 |
|
44 |
def calculate_diversity(text):
|
|
|
|
|
|
|
45 |
|
46 |
-
|
47 |
-
for i in string.punctuation:
|
48 |
-
stop_words.add(i)
|
49 |
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
return 1,"More Text Required"
|
55 |
|
|
|
|
|
|
|
|
|
56 |
|
|
|
57 |
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
62 |
|
63 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
64 |
|
65 |
-
for pos, comp_word in enumerate(tokenized_text):
|
66 |
-
if anc_word in sim_words.get(pos, []):
|
67 |
-
if anc_word == sim_words[pos][0]:
|
68 |
-
sim_words[idx] = sim_words[pos]
|
69 |
-
continue
|
70 |
-
try:
|
71 |
-
if not comp_word in stop_words and cosine_similarity(w2v[anc_word].reshape(1, -1), w2v[comp_word].reshape(1, -1)) > .75:
|
72 |
-
vocab.append(comp_word)
|
73 |
-
sim_words[idx] = vocab
|
74 |
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
scores = {}
|
80 |
-
for key, value in sim_words.items():
|
81 |
-
if len(value) == 1:
|
82 |
-
scores[key] = 1
|
83 |
-
continue
|
84 |
-
if len(value) == 2:
|
85 |
-
scores[key] = -1
|
86 |
-
continue
|
87 |
-
t_sim = len(value) - 1
|
88 |
-
t_rep = (len(value) - 1) - (len(set(value[1:])))
|
89 |
-
|
90 |
-
score = ((t_sim - t_rep)/t_sim)**2
|
91 |
-
|
92 |
-
scores[key] = score
|
93 |
-
|
94 |
-
mean_score = 0
|
95 |
-
total = 0
|
96 |
-
|
97 |
-
for value in scores.values():
|
98 |
-
if value == -1:
|
99 |
-
continue
|
100 |
-
mean_score += value
|
101 |
-
total += 1
|
102 |
-
|
103 |
-
return scores, mean_score/total
|
104 |
|
105 |
|
106 |
def dict_to_list(dictionary, max_size=10):
|
107 |
outer_list = []
|
108 |
inner_list = []
|
109 |
|
110 |
-
for value in dictionary.
|
111 |
inner_list.append(value)
|
112 |
if len(inner_list) == max_size:
|
113 |
outer_list.append(inner_list)
|
@@ -140,7 +162,7 @@ def stats(text):
|
|
140 |
|
141 |
def predict(text, tokenizer=tokenizer):
|
142 |
model.eval()
|
143 |
-
model.to(
|
144 |
|
145 |
def prepare_data(text, tokenizer):
|
146 |
input_ids = []
|
@@ -166,14 +188,25 @@ def predict(text, tokenizer=tokenizer):
|
|
166 |
tokenized_example_text = prepare_data(text, tokenizer)
|
167 |
with torch.no_grad():
|
168 |
result = model(
|
169 |
-
tokenized_example_text['input_ids'].to(
|
170 |
-
attention_mask=tokenized_example_text['attention_masks'].to(
|
171 |
return_dict=True
|
172 |
).logits
|
173 |
|
174 |
return result
|
175 |
|
176 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
177 |
def reading_difficulty(excerpt):
|
178 |
if len(excerpt) == 0:
|
179 |
return "No Text Provided"
|
@@ -192,12 +225,14 @@ def reading_difficulty(excerpt):
|
|
192 |
win_preds.append(predict(text, tokenizer).item())
|
193 |
result = statistics.mean(win_preds)
|
194 |
score = -(result * 1.786 + 6.4) + 10
|
195 |
-
return score
|
|
|
196 |
|
197 |
else:
|
198 |
result = predict(excerpt).item()
|
199 |
score = -(result * 1.786 + 6.4) + 10
|
200 |
-
return score
|
|
|
201 |
|
202 |
|
203 |
def calculate_stats(file_name, data_index):
|
@@ -237,11 +272,11 @@ def calculate_stats(file_name, data_index):
|
|
237 |
def transcribe(audio):
|
238 |
# speech to text using pipeline
|
239 |
text = p(audio)["text"]
|
240 |
-
transcription.append(text)
|
241 |
return text
|
242 |
|
243 |
|
244 |
def compute_score(target, actual):
|
|
|
245 |
target = target.lower()
|
246 |
actual = actual.lower()
|
247 |
return fuzz.ratio(target, actual)
|
@@ -256,94 +291,164 @@ def phon(text):
|
|
256 |
pronun.append(alph[word][0])
|
257 |
except Exception as e:
|
258 |
pronun.append(word)
|
259 |
-
return pronun
|
260 |
|
|
|
|
|
|
|
|
|
|
|
261 |
|
262 |
-
|
263 |
-
|
264 |
-
|
265 |
-
|
266 |
-
|
267 |
-
if actual_audio is not None:
|
268 |
-
actual = p(actual_audio)["text"]
|
269 |
-
speech_score = compute_score(target, actual)
|
270 |
-
|
271 |
-
return "Difficulty Score: " + str(reading_difficulty(actual)), "Transcript: " + str(
|
272 |
-
actual.lower()), "Diversity Score: " + str(calculate_diversity(target)[1]), "Speech Score: " + str(speech_score)
|
273 |
-
div = calculate_diversity(text)
|
274 |
-
transcription = []
|
275 |
-
if audio is not None:
|
276 |
-
text = p(audio)["text"]
|
277 |
-
transcription.append(text)
|
278 |
-
state = div[0]
|
279 |
-
return "Difficulty Score: " + str(reading_difficulty(text)), "Transcript: " + str(
|
280 |
-
transcription[-1].lower()), "Diversity Score: " + str(div[1]), "No Inputs"
|
281 |
-
|
282 |
-
return "Difficulty Score: " + str(reading_difficulty(text)), "Diversity Score: " + str(
|
283 |
-
div[1]), "No Audio Provided", "No Audio Provided"
|
284 |
|
285 |
|
286 |
def plot():
|
287 |
-
text = state
|
288 |
diversity = calculate_diversity(text)[0]
|
|
|
289 |
df = pd.DataFrame(dict_to_list(diversity))
|
290 |
return heatmap(diversity, df)
|
291 |
|
292 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
293 |
|
294 |
-
|
295 |
-
|
296 |
-
|
297 |
-
|
298 |
-
|
299 |
-
|
300 |
-
|
301 |
-
|
302 |
-
if
|
303 |
-
|
304 |
-
|
305 |
-
|
306 |
-
|
307 |
-
|
308 |
-
|
309 |
-
|
310 |
-
|
311 |
-
|
312 |
-
|
313 |
-
|
314 |
-
|
315 |
-
|
316 |
-
|
317 |
-
|
318 |
-
)
|
319 |
-
|
320 |
-
|
321 |
-
|
322 |
-
|
323 |
-
|
324 |
-
|
325 |
-
|
326 |
-
|
327 |
-
|
328 |
-
|
329 |
-
|
330 |
-
|
331 |
-
|
332 |
-
|
333 |
-
|
334 |
-
|
335 |
-
|
336 |
-
|
337 |
-
|
338 |
-
|
339 |
-
|
340 |
-
|
341 |
-
|
342 |
-
|
343 |
-
|
344 |
-
|
345 |
-
|
346 |
-
|
347 |
-
|
348 |
-
|
349 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import csv
|
|
|
2 |
import string
|
3 |
|
4 |
import gensim.downloader as api
|
5 |
+
import matplotlib.pyplot as plt
|
6 |
import nltk
|
7 |
+
import numpy as np
|
8 |
import pandas as pd
|
9 |
import readability
|
10 |
import seaborn as sns
|
11 |
import torch
|
12 |
from fuzzywuzzy import fuzz
|
13 |
from nltk.corpus import stopwords
|
14 |
+
from nltk.corpus import wordnet as wn
|
15 |
from nltk.tokenize import word_tokenize
|
16 |
from sklearn.metrics.pairwise import cosine_similarity
|
17 |
from transformers import DistilBertTokenizer
|
18 |
from transformers import pipeline
|
19 |
|
20 |
+
nltk.download('wordnet')
|
21 |
+
|
22 |
+
nltk.download('omw-1.4')
|
23 |
+
|
24 |
nltk.download('cmudict')
|
25 |
|
26 |
nltk.download('stopwords')
|
|
|
34 |
|
35 |
# loading model
|
36 |
PATH = 'pytorchBERTmodel'
|
37 |
+
model = torch.load(PATH)
|
38 |
model.eval()
|
39 |
|
40 |
+
model.to(device)
|
41 |
|
42 |
p = pipeline("automatic-speech-recognition")
|
43 |
|
44 |
+
|
45 |
+
def syns(word):
|
46 |
+
synonyms = []
|
47 |
+
for syn in wn.synsets(word):
|
48 |
+
for lm in syn.lemmas():
|
49 |
+
synonyms.append(lm.name())
|
50 |
+
return set(synonyms)
|
51 |
+
|
52 |
+
|
53 |
w2v = dict({})
|
54 |
+
for idx, key in enumerate(glove_vectors.wv.vocab):
|
55 |
+
w2v[key] = glove_vectors.wv.get_vector(key)
|
56 |
|
57 |
|
58 |
def calculate_diversity(text):
|
59 |
+
stop_words = set(stopwords.words('english'))
|
60 |
+
for i in string.punctuation:
|
61 |
+
stop_words.add(i)
|
62 |
|
63 |
+
tokenized_text = word_tokenize(text)
|
|
|
|
|
64 |
|
65 |
+
tokenized_text = list(map(lambda word: word.lower(), tokenized_text))
|
66 |
+
sim_words = {}
|
67 |
+
if len(tokenized_text) <= 1:
|
68 |
+
return 1, "More Text Required"
|
|
|
69 |
|
70 |
+
for idx, anc in enumerate(tokenized_text):
|
71 |
+
if anc in stop_words or not anc in w2v or anc.isdigit():
|
72 |
+
sim_words[idx] = '@'
|
73 |
+
continue
|
74 |
|
75 |
+
vocab = [anc]
|
76 |
|
77 |
+
for pos, comp in enumerate(tokenized_text):
|
78 |
+
if pos == idx:
|
79 |
+
continue
|
80 |
+
if comp in stop_words:
|
81 |
+
continue
|
82 |
+
if not comp.isalpha():
|
83 |
+
continue
|
84 |
+
try:
|
85 |
+
if cosine_similarity(w2v[anc].reshape(1, -1), w2v[comp].reshape(1, -1)) > .7 or comp in syns(anc):
|
86 |
+
vocab.append(comp)
|
87 |
+
except KeyError:
|
88 |
+
continue
|
89 |
+
sim_words[idx] = vocab
|
90 |
+
print(sim_words)
|
91 |
+
scores = {}
|
92 |
+
for key, value in sim_words.items():
|
93 |
+
if len(value) == 1:
|
94 |
+
scores[key] = -1
|
95 |
+
continue
|
96 |
+
# if len(value) == 2:
|
97 |
+
# scores[key] = -1
|
98 |
+
# continue
|
99 |
+
t_sim = len(value)
|
100 |
+
t_rep = (len(value)) - (len(set(value)))
|
101 |
+
|
102 |
+
score = ((t_sim - t_rep) / t_sim) ** 2
|
103 |
+
|
104 |
+
scores[key] = score
|
105 |
+
|
106 |
+
mean_score = 0
|
107 |
+
total = 0
|
108 |
|
109 |
+
for value in scores.values():
|
110 |
+
if value == -1:
|
111 |
+
continue
|
112 |
+
mean_score += value
|
113 |
+
total += 1
|
114 |
+
try:
|
115 |
+
return scores, {"Diversity Score": mean_score / total}
|
116 |
+
except ZeroDivisionError:
|
117 |
+
return scores, {"Dviersity Score": "Not Enough Data"}
|
118 |
+
|
119 |
+
|
120 |
+
def get_scores(text):
|
121 |
+
return calculate_diversity(text)[0]
|
122 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
123 |
|
124 |
+
def get_mean_score(text):
|
125 |
+
return calculate_diversity(text)[1]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
126 |
|
127 |
|
128 |
def dict_to_list(dictionary, max_size=10):
|
129 |
outer_list = []
|
130 |
inner_list = []
|
131 |
|
132 |
+
for key, value in dictionary.items():
|
133 |
inner_list.append(value)
|
134 |
if len(inner_list) == max_size:
|
135 |
outer_list.append(inner_list)
|
|
|
162 |
|
163 |
def predict(text, tokenizer=tokenizer):
|
164 |
model.eval()
|
165 |
+
model.to(device)
|
166 |
|
167 |
def prepare_data(text, tokenizer):
|
168 |
input_ids = []
|
|
|
188 |
tokenized_example_text = prepare_data(text, tokenizer)
|
189 |
with torch.no_grad():
|
190 |
result = model(
|
191 |
+
tokenized_example_text['input_ids'].to(device),
|
192 |
+
attention_mask=tokenized_example_text['attention_masks'].to(device),
|
193 |
return_dict=True
|
194 |
).logits
|
195 |
|
196 |
return result
|
197 |
|
198 |
|
199 |
+
def level(score):
|
200 |
+
if score <= 3:
|
201 |
+
return "Elementary School"
|
202 |
+
elif 3 <= score <= 6:
|
203 |
+
return "Middle School"
|
204 |
+
elif 6 <= score <= 8:
|
205 |
+
return "High School"
|
206 |
+
else:
|
207 |
+
return "College"
|
208 |
+
|
209 |
+
|
210 |
def reading_difficulty(excerpt):
|
211 |
if len(excerpt) == 0:
|
212 |
return "No Text Provided"
|
|
|
225 |
win_preds.append(predict(text, tokenizer).item())
|
226 |
result = statistics.mean(win_preds)
|
227 |
score = -(result * 1.786 + 6.4) + 10
|
228 |
+
return "Difficulty Level: " + str(round(score, 2)) + '/10' + ' | A ' + str(
|
229 |
+
level(score)) + " student could understand this"
|
230 |
|
231 |
else:
|
232 |
result = predict(excerpt).item()
|
233 |
score = -(result * 1.786 + 6.4) + 10
|
234 |
+
return 'Difficulty Level: ' + str(round(score, 2)) + '/10' + ' | A ' + str(
|
235 |
+
level(score)) + " student could understand this"
|
236 |
|
237 |
|
238 |
def calculate_stats(file_name, data_index):
|
|
|
272 |
def transcribe(audio):
|
273 |
# speech to text using pipeline
|
274 |
text = p(audio)["text"]
|
|
|
275 |
return text
|
276 |
|
277 |
|
278 |
def compute_score(target, actual):
|
279 |
+
print(target)
|
280 |
target = target.lower()
|
281 |
actual = actual.lower()
|
282 |
return fuzz.ratio(target, actual)
|
|
|
291 |
pronun.append(alph[word][0])
|
292 |
except Exception as e:
|
293 |
pronun.append(word)
|
|
|
294 |
|
295 |
+
def remove_digits(lists):
|
296 |
+
for lst in lists:
|
297 |
+
for idx, word in enumerate(lst):
|
298 |
+
lst[idx] = ''.join([letter for letter in word if not letter.isdigit()])
|
299 |
+
return lists
|
300 |
|
301 |
+
output = []
|
302 |
+
for i in remove_digits(pronun):
|
303 |
+
output.append('-'.join(i).lower())
|
304 |
+
return ' '.join(output)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
305 |
|
306 |
|
307 |
def plot():
|
|
|
308 |
diversity = calculate_diversity(text)[0]
|
309 |
+
print(diversity)
|
310 |
df = pd.DataFrame(dict_to_list(diversity))
|
311 |
return heatmap(diversity, df)
|
312 |
|
313 |
|
314 |
+
def diversity_inter(text):
|
315 |
+
words = word_tokenize(text)
|
316 |
+
scores = get_scores(text)
|
317 |
+
interpret_values = [('', 0.0)]
|
318 |
+
for key, value in scores.items():
|
319 |
+
interpret_values.append((words[key], value))
|
320 |
+
interpret_values.append(('', 0.0))
|
321 |
+
print(interpret_values)
|
322 |
+
return {'original': text, 'interpretation': interpret_values}
|
323 |
|
324 |
+
|
325 |
+
def sliding_window(text):
|
326 |
+
wind_preds = []
|
327 |
+
windows = []
|
328 |
+
new_values = []
|
329 |
+
heat_map = []
|
330 |
+
words = word_tokenize(text)
|
331 |
+
for idx, text in enumerate(words):
|
332 |
+
if idx <= len(words) - 26:
|
333 |
+
x = ' '.join(words[idx: idx + 25])
|
334 |
+
windows.append(x)
|
335 |
+
|
336 |
+
for text in windows:
|
337 |
+
prediction = -(predict(text).item() * 1.786 + 6.4) + 10
|
338 |
+
wind_preds.append(prediction)
|
339 |
+
|
340 |
+
size = 25
|
341 |
+
for i in wind_preds:
|
342 |
+
for j in range(size):
|
343 |
+
new_values.append(i)
|
344 |
+
|
345 |
+
heat_map = []
|
346 |
+
for idx, i in enumerate(new_values):
|
347 |
+
window = new_values[idx:idx + size]
|
348 |
+
heat_map.append(np.mean(window))
|
349 |
+
compressed_map = []
|
350 |
+
for idx, i in enumerate(heat_map):
|
351 |
+
if idx % size == 0:
|
352 |
+
window = heat_map[idx:idx + size]
|
353 |
+
compressed_map.append(np.mean(window))
|
354 |
+
|
355 |
+
inter_scores = compressed_map
|
356 |
+
while len(inter_scores) <= len(words) - 1:
|
357 |
+
inter_scores.append(compressed_map[-1])
|
358 |
+
|
359 |
+
x = list(range(len(inter_scores)))
|
360 |
+
y = inter_scores
|
361 |
+
|
362 |
+
fig, ax = plt.subplots()
|
363 |
+
|
364 |
+
ax.plot(x, y, color='orange', linewidth=2)
|
365 |
+
ax.grid(False)
|
366 |
+
plt.xlabel('Word Number', fontweight='bold')
|
367 |
+
plt.ylabel('Difficulty Score', fontweight='bold')
|
368 |
+
fig.patch.set_facecolor('white')
|
369 |
+
plt.suptitle('Difficulty Score Across Text', fontsize=14, fontweight='bold')
|
370 |
+
plt.style.use('ggplot')
|
371 |
+
|
372 |
+
fig = plt.gcf()
|
373 |
+
|
374 |
+
map = [('', 0)]
|
375 |
+
maxy = max(inter_scores)
|
376 |
+
miny = min(inter_scores)
|
377 |
+
spread = maxy - miny
|
378 |
+
|
379 |
+
for idx, i in enumerate(words):
|
380 |
+
map.append((i, (inter_scores[idx] - miny) / spread))
|
381 |
+
map.append(('', 0))
|
382 |
+
|
383 |
+
return fig, map
|
384 |
+
|
385 |
+
|
386 |
+
def get_plot(text):
|
387 |
+
return sliding_window(text)[0]
|
388 |
+
|
389 |
+
|
390 |
+
def get_dif_inter(text):
|
391 |
+
return {'original': text, 'interpretation': sliding_window(text)[1]}
|
392 |
+
|
393 |
+
|
394 |
+
def speech_to_text(speech, target):
|
395 |
+
text = p(speech)["text"]
|
396 |
+
return text.lower(), {'Pronunciation Score': compute_score(text, target) / 100}, phon(target)
|
397 |
+
|
398 |
+
|
399 |
+
def my_i_func(text):
|
400 |
+
return {"original": "", "interpretation": [('', 0.0), ('what', -0.2), ('great', 0.3), ('day', 0.5), ('', 0.0)]}
|
401 |
+
|
402 |
+
|
403 |
+
inter = {"original": "what a wonderful day", "interpretation": [0, .2, .3, .5]}
|
404 |
+
|
405 |
+
with gr.Blocks() as demo:
|
406 |
+
with gr.Column():
|
407 |
+
with gr.Row():
|
408 |
+
with gr.Box():
|
409 |
+
with gr.Column():
|
410 |
+
with gr.Group():
|
411 |
+
with gr.Tabs():
|
412 |
+
with gr.TabItem("Text"):
|
413 |
+
in_text = gr.Textbox(label="In Text")
|
414 |
+
grade = gr.Button("Grade Your Text")
|
415 |
+
with gr.TabItem("Speech"):
|
416 |
+
audio_file = gr.Audio(source="microphone", type="filepath")
|
417 |
+
grade1 = gr.Button("Grade Your Speech")
|
418 |
+
|
419 |
+
with gr.Box():
|
420 |
+
diff_output = gr.Label(label='Difficulty Level', show_label=True)
|
421 |
+
gr.Markdown("Diversity Score Across Text")
|
422 |
+
plotter = gr.Plot()
|
423 |
+
|
424 |
+
with gr.Row():
|
425 |
+
with gr.Box():
|
426 |
+
div_output = gr.Label(label='Diversity Score', show_label=False)
|
427 |
+
gr.Markdown("Diversity Heamap")
|
428 |
+
interpretation = gr.components.Interpretation(in_text, label="Diversity Heapmap")
|
429 |
+
# gr.DataFrame(df)
|
430 |
+
with gr.Box():
|
431 |
+
# plotter = gr.Plot()
|
432 |
+
# gr.Markdown("*Nominal Score May Not Represent ")
|
433 |
+
interpretation2 = gr.components.Interpretation(in_text, label="Difficulty Heapmap")
|
434 |
+
with gr.Row():
|
435 |
+
with gr.Box():
|
436 |
+
with gr.Group():
|
437 |
+
target = gr.Textbox(label="Target Text")
|
438 |
+
with gr.Group():
|
439 |
+
audio_file1 = gr.Audio(source="microphone", type="filepath")
|
440 |
+
b1 = gr.Button("Grade Your Pronunciation")
|
441 |
+
with gr.Box():
|
442 |
+
some_val = gr.Label()
|
443 |
+
text = gr.Textbox()
|
444 |
+
phones = gr.Textbox()
|
445 |
+
|
446 |
+
grade.click(reading_difficulty, inputs=in_text, outputs=diff_output)
|
447 |
+
grade.click(get_mean_score, inputs=in_text, outputs=div_output)
|
448 |
+
grade.click(diversity_inter, inputs=in_text, outputs=interpretation)
|
449 |
+
grade.click(get_dif_inter, inputs=in_text, outputs=interpretation2)
|
450 |
+
grade.click(get_plot, inputs=in_text, outputs=plotter)
|
451 |
+
# grade1.click(transcribe, inputs=input_audio, outputs=in_text)
|
452 |
+
# pronun.click(transcribe, inputs=pronon_audio, outputs=trans)
|
453 |
+
b1.click(speech_to_text, inputs=[audio_file1, target], outputs=[text, some_val, phones])
|
454 |
+
demo.launch(debug=True)
|