Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -3,18 +3,21 @@ import torch
|
|
3 |
from transformers import AutoTokenizer, T5ForConditionalGeneration, pipeline
|
4 |
from sentence_transformers import SentenceTransformer, util
|
5 |
import requests
|
6 |
-
import warnings
|
7 |
import os
|
8 |
-
|
|
|
9 |
|
10 |
-
#
|
11 |
-
|
12 |
-
warnings.filterwarnings("ignore", category=
|
13 |
-
warnings.filterwarnings("ignore"
|
|
|
14 |
|
15 |
-
|
|
|
|
|
16 |
|
17 |
-
#
|
18 |
def segment_into_sentences_groq(passage):
|
19 |
headers = {
|
20 |
"Authorization": f"Bearer {GROQ_API_KEY}",
|
@@ -25,146 +28,109 @@ def segment_into_sentences_groq(passage):
|
|
25 |
"messages": [
|
26 |
{
|
27 |
"role": "system",
|
28 |
-
"content": "
|
29 |
},
|
30 |
{
|
31 |
"role": "user",
|
32 |
-
"content": f"Segment
|
33 |
}
|
34 |
],
|
35 |
-
"temperature": 0
|
36 |
-
"max_tokens":
|
37 |
}
|
38 |
-
|
39 |
response = requests.post("https://api.groq.com/openai/v1/chat/completions", json=payload, headers=headers)
|
40 |
if response.status_code == 200:
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
except (KeyError, IndexError):
|
46 |
-
raise ValueError("Unexpected response structure from Groq API.")
|
47 |
else:
|
48 |
raise ValueError(f"Groq API error: {response.text}")
|
49 |
|
50 |
-
|
51 |
class TextEnhancer:
|
52 |
def __init__(self):
|
53 |
self.device = "cuda" if torch.cuda.is_available() else "cpu"
|
54 |
-
self.executor = ThreadPoolExecutor(max_workers=3) # Parallel processing pool
|
55 |
-
|
56 |
-
# Load models
|
57 |
-
self._load_models()
|
58 |
-
|
59 |
-
def _load_models(self):
|
60 |
self.paraphrase_tokenizer = AutoTokenizer.from_pretrained("prithivida/parrot_paraphraser_on_T5")
|
61 |
self.paraphrase_model = T5ForConditionalGeneration.from_pretrained("prithivida/parrot_paraphraser_on_T5").to(self.device)
|
62 |
-
|
63 |
self.grammar_pipeline = pipeline(
|
64 |
"text2text-generation",
|
65 |
model="Grammarly/coedit-large",
|
66 |
device=0 if self.device == "cuda" else -1
|
67 |
)
|
68 |
-
|
69 |
self.similarity_model = SentenceTransformer('paraphrase-MiniLM-L6-v2').to(self.device)
|
70 |
|
71 |
-
def enhance_text(self, text, min_similarity=0.8):
|
72 |
sentences = segment_into_sentences_groq(text)
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
96 |
-
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
|
107 |
-
|
108 |
-
|
109 |
-
|
110 |
-
|
111 |
-
|
112 |
-
|
113 |
-
|
114 |
-
|
115 |
-
|
116 |
-
|
117 |
-
|
118 |
-
|
119 |
-
|
120 |
-
|
121 |
-
|
122 |
-
else:
|
123 |
-
return sentence
|
124 |
-
|
125 |
-
def _humanize_text(self, text):
|
126 |
-
# Introduce minor variations to mimic human-written text
|
127 |
-
import random
|
128 |
-
contractions = {"can't": "cannot", "won't": "will not", "it's": "it is"}
|
129 |
-
words = text.split()
|
130 |
-
text = " ".join([contractions.get(word, word) if random.random() > 0.9 else word for word in words])
|
131 |
-
|
132 |
-
if random.random() > 0.7:
|
133 |
-
text = text.replace(" and ", ", and ")
|
134 |
-
return text
|
135 |
-
|
136 |
-
|
137 |
def create_interface():
|
138 |
enhancer = TextEnhancer()
|
139 |
-
|
140 |
def process_text(text, similarity_threshold):
|
141 |
try:
|
142 |
return enhancer.enhance_text(text, min_similarity=similarity_threshold / 100)
|
143 |
except Exception as e:
|
144 |
return f"Error: {str(e)}"
|
145 |
-
|
146 |
-
|
147 |
fn=process_text,
|
148 |
inputs=[
|
149 |
-
gr.Textbox(
|
150 |
-
|
151 |
-
placeholder="Enter text to enhance...",
|
152 |
-
lines=10
|
153 |
-
),
|
154 |
-
gr.Slider(
|
155 |
-
minimum=50,
|
156 |
-
maximum=100,
|
157 |
-
value=80,
|
158 |
-
label="Minimum Semantic Similarity (%)"
|
159 |
-
)
|
160 |
],
|
161 |
-
outputs=gr.Textbox(label="Enhanced Text"
|
162 |
title="Text Enhancement System",
|
163 |
-
description="
|
164 |
)
|
165 |
-
return interface
|
166 |
-
|
167 |
|
168 |
if __name__ == "__main__":
|
169 |
interface = create_interface()
|
170 |
-
interface.launch()
|
|
|
3 |
from transformers import AutoTokenizer, T5ForConditionalGeneration, pipeline
|
4 |
from sentence_transformers import SentenceTransformer, util
|
5 |
import requests
|
|
|
6 |
import os
|
7 |
+
import warnings
|
8 |
+
from transformers import logging
|
9 |
|
10 |
+
# Suppress warnings
|
11 |
+
warnings.filterwarnings("ignore", category=FutureWarning)
|
12 |
+
warnings.filterwarnings("ignore", category=UserWarning)
|
13 |
+
warnings.filterwarnings("ignore")
|
14 |
+
logging.set_verbosity_error()
|
15 |
|
16 |
+
# Set API keys and environment variables
|
17 |
+
GROQ_API_KEY = os.getenv("GROQ_API_KEY") # Ensure you set this in Hugging Face Spaces
|
18 |
+
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
|
19 |
|
20 |
+
# Groq API sentence segmentation
|
21 |
def segment_into_sentences_groq(passage):
|
22 |
headers = {
|
23 |
"Authorization": f"Bearer {GROQ_API_KEY}",
|
|
|
28 |
"messages": [
|
29 |
{
|
30 |
"role": "system",
|
31 |
+
"content": "Segment sentences by adding '1!2@3#' at the end of each sentence."
|
32 |
},
|
33 |
{
|
34 |
"role": "user",
|
35 |
+
"content": f"Segment the passage: {passage}"
|
36 |
}
|
37 |
],
|
38 |
+
"temperature": 1.0,
|
39 |
+
"max_tokens": 8192
|
40 |
}
|
|
|
41 |
response = requests.post("https://api.groq.com/openai/v1/chat/completions", json=payload, headers=headers)
|
42 |
if response.status_code == 200:
|
43 |
+
data = response.json()
|
44 |
+
segmented_text = data.get("choices", [{}])[0].get("message", {}).get("content", "")
|
45 |
+
sentences = segmented_text.split("1!2@3#")
|
46 |
+
return [sentence.strip() for sentence in sentences if sentence.strip()]
|
|
|
|
|
47 |
else:
|
48 |
raise ValueError(f"Groq API error: {response.text}")
|
49 |
|
50 |
+
# Text enhancement class
|
51 |
class TextEnhancer:
|
52 |
def __init__(self):
|
53 |
self.device = "cuda" if torch.cuda.is_available() else "cpu"
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
self.paraphrase_tokenizer = AutoTokenizer.from_pretrained("prithivida/parrot_paraphraser_on_T5")
|
55 |
self.paraphrase_model = T5ForConditionalGeneration.from_pretrained("prithivida/parrot_paraphraser_on_T5").to(self.device)
|
|
|
56 |
self.grammar_pipeline = pipeline(
|
57 |
"text2text-generation",
|
58 |
model="Grammarly/coedit-large",
|
59 |
device=0 if self.device == "cuda" else -1
|
60 |
)
|
|
|
61 |
self.similarity_model = SentenceTransformer('paraphrase-MiniLM-L6-v2').to(self.device)
|
62 |
|
63 |
+
def enhance_text(self, text, min_similarity=0.8, max_variations=2):
|
64 |
sentences = segment_into_sentences_groq(text)
|
65 |
+
enhanced_sentences = []
|
66 |
+
|
67 |
+
for sentence in sentences:
|
68 |
+
if not sentence.strip():
|
69 |
+
continue
|
70 |
+
|
71 |
+
# Generate paraphrases
|
72 |
+
inputs = self.paraphrase_tokenizer(
|
73 |
+
f"paraphrase: {sentence}",
|
74 |
+
return_tensors="pt",
|
75 |
+
padding=True,
|
76 |
+
max_length=150,
|
77 |
+
truncation=True
|
78 |
+
).to(self.device)
|
79 |
+
|
80 |
+
outputs = self.paraphrase_model.generate(
|
81 |
+
**inputs,
|
82 |
+
max_length=150,
|
83 |
+
num_return_sequences=max_variations,
|
84 |
+
num_beams=max_variations
|
85 |
+
)
|
86 |
+
paraphrases = [
|
87 |
+
self.paraphrase_tokenizer.decode(output, skip_special_tokens=True)
|
88 |
+
for output in outputs
|
89 |
+
]
|
90 |
+
|
91 |
+
# Calculate semantic similarity
|
92 |
+
sentence_embedding = self.similarity_model.encode(sentence)
|
93 |
+
paraphrase_embeddings = self.similarity_model.encode(paraphrases)
|
94 |
+
similarities = util.cos_sim(sentence_embedding, paraphrase_embeddings)
|
95 |
+
|
96 |
+
# Select the most similar paraphrase
|
97 |
+
valid_paraphrases = [
|
98 |
+
para for para, sim in zip(paraphrases, similarities[0])
|
99 |
+
if sim >= min_similarity
|
100 |
+
]
|
101 |
+
if valid_paraphrases:
|
102 |
+
corrected = self.grammar_pipeline(
|
103 |
+
valid_paraphrases[0],
|
104 |
+
max_length=150,
|
105 |
+
num_return_sequences=1
|
106 |
+
)[0]["generated_text"]
|
107 |
+
enhanced_sentences.append(corrected)
|
108 |
+
else:
|
109 |
+
enhanced_sentences.append(sentence)
|
110 |
+
|
111 |
+
return ". ".join(enhanced_sentences).strip() + "."
|
112 |
+
|
113 |
+
# Gradio interface
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
114 |
def create_interface():
|
115 |
enhancer = TextEnhancer()
|
116 |
+
|
117 |
def process_text(text, similarity_threshold):
|
118 |
try:
|
119 |
return enhancer.enhance_text(text, min_similarity=similarity_threshold / 100)
|
120 |
except Exception as e:
|
121 |
return f"Error: {str(e)}"
|
122 |
+
|
123 |
+
return gr.Interface(
|
124 |
fn=process_text,
|
125 |
inputs=[
|
126 |
+
gr.Textbox(lines=10, placeholder="Enter text to enhance...", label="Input Text"),
|
127 |
+
gr.Slider(50, 100, 80, label="Minimum Semantic Similarity (%)")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
128 |
],
|
129 |
+
outputs=gr.Textbox(lines=10, label="Enhanced Text"),
|
130 |
title="Text Enhancement System",
|
131 |
+
description="Enhance text quality with semantic preservation."
|
132 |
)
|
|
|
|
|
133 |
|
134 |
if __name__ == "__main__":
|
135 |
interface = create_interface()
|
136 |
+
interface.launch(server_name="0.0.0.0", server_port=7860)
|