Spaces:
Sleeping
Sleeping
Update src/streamlit_app.py
Browse files- src/streamlit_app.py +302 -308
src/streamlit_app.py
CHANGED
@@ -21,247 +21,241 @@ import torch # Import torch
|
|
21 |
# nltk.download('averaged_perceptron_tagger', download_dir=nltk_data_dir, quiet=True)
|
22 |
# nltk.data.path.append(nltk_data_dir)
|
23 |
|
24 |
-
nltk_data_dir = "/tmp/nltk_data"
|
25 |
-
os.makedirs(nltk_data_dir, exist_ok=True)
|
26 |
-
nltk.data.path.append(nltk_data_dir)
|
27 |
-
|
28 |
-
# nltk.download('punkt', download_dir=nltk_data_dir)
|
29 |
-
# nltk.download('averaged_perceptron_tagger', download_dir=nltk_data_dir)
|
30 |
-
nltk.download('averaged_perceptron_tagger', download_dir=nltk_data_dir, quiet=True)
|
31 |
-
nltk.download('punkt', download_dir=nltk_data_dir, quiet=True)
|
32 |
-
nltk.download('punkt_tab', download_dir=nltk_data_dir, quiet=True)
|
33 |
-
|
34 |
-
|
35 |
-
# Load models
|
36 |
-
# summarizer = pipeline("summarization", model="sshleifer/distilbart-cnn-12-6", device=torch.device("cpu"))
|
37 |
-
# summarizer = pipeline(
|
38 |
-
# "summarization",
|
39 |
-
# model="sshleifer/distilbart-cnn-12-6",
|
40 |
-
# from_flax=True,
|
41 |
-
# device=-1 # CPU mode, use device=0 for GPU
|
42 |
-
# )
|
43 |
-
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
|
44 |
-
qa_pipeline = pipeline("question-answering", model="distilbert-base-cased-distilled-squad", device=torch.device("cpu"))
|
45 |
-
embedding_model = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
|
46 |
-
vector_dbs = {} # Dictionary to store multiple vector databases, keyed by document title
|
47 |
-
extracted_texts = {} # Dictionary to store extracted text, keyed by document title
|
48 |
-
current_doc_title = None
|
49 |
-
|
50 |
-
# ---------------------------------------------
|
51 |
-
# Extract text from PDF or Image
|
52 |
-
# ---------------------------------------------
|
53 |
-
def extract_text(uploaded_file):
|
54 |
-
global current_doc_title
|
55 |
-
current_doc_title = uploaded_file.name
|
56 |
-
suffix = uploaded_file.name.lower()
|
57 |
-
with tempfile.NamedTemporaryFile(delete=False) as tmp:
|
58 |
-
tmp.write(uploaded_file.read())
|
59 |
-
path = tmp.name
|
60 |
-
|
61 |
-
text = ""
|
62 |
-
if suffix.endswith(".pdf"):
|
63 |
-
with pdfplumber.open(path) as pdf:
|
64 |
-
for page in pdf.pages:
|
65 |
-
page_text = page.extract_text()
|
66 |
-
if page_text:
|
67 |
-
text += page_text + "\n"
|
68 |
-
else:
|
69 |
-
try:
|
70 |
-
text = pytesseract.image_to_string(Image.open(path))
|
71 |
-
except Exception as e:
|
72 |
-
st.error(f"Error during OCR for {uploaded_file.name}: {e}")
|
73 |
-
text = ""
|
74 |
-
|
75 |
-
os.remove(path)
|
76 |
-
return text.strip()
|
77 |
-
|
78 |
-
# ---------------------------------------------
|
79 |
-
# Store Embeddings in FAISS
|
80 |
-
# ---------------------------------------------
|
81 |
-
def store_vector(text):
|
82 |
-
global vector_dbs, current_doc_title
|
83 |
-
text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
|
84 |
-
docs = text_splitter.create_documents([text])
|
85 |
-
for doc in docs:
|
86 |
-
doc.metadata = {"title": current_doc_title} # Add document title as metadata
|
87 |
-
|
88 |
-
if current_doc_title in vector_dbs:
|
89 |
-
vector_dbs[current_doc_title].add_documents(docs) # Append to existing DB
|
90 |
-
else:
|
91 |
-
vector_dbs[current_doc_title] = FAISS.from_documents(docs, embedding_model)
|
92 |
-
|
93 |
-
# ---------------------------------------------
|
94 |
-
# Summarize Text
|
95 |
-
# ---------------------------------------------
|
96 |
-
def summarize(text):
|
97 |
-
if len(text.split()) < 100:
|
98 |
-
return "Text too short to summarize."
|
99 |
-
chunks = [text[i : i + 1024] for i in range(0, len(text), 1024)]
|
100 |
-
summaries = []
|
101 |
-
for chunk in chunks:
|
102 |
-
max_len = int(len(chunk.split()) * 0.6)
|
103 |
-
max_len = max(30, min(max_len, 150))
|
104 |
-
try:
|
105 |
-
summary = summarizer(chunk, max_length=max_len, min_length=20)[0]["summary_text"]
|
106 |
-
summaries.append(summary)
|
107 |
-
except Exception as e:
|
108 |
-
st.error(f"Error during summarization: {e}")
|
109 |
-
return "An error occurred during summarization."
|
110 |
-
return " ".join(summaries)
|
111 |
-
|
112 |
-
# ---------------------------------------------
|
113 |
-
# Question Answering
|
114 |
-
# ---------------------------------------------
|
115 |
-
def topic_search(question, doc_title=None):
|
116 |
-
global vector_dbs
|
117 |
-
if not vector_dbs:
|
118 |
-
st.warning("Please upload and process a file first in the 'Upload & Extract' tab.")
|
119 |
-
return ""
|
120 |
-
|
121 |
-
try:
|
122 |
-
if doc_title and doc_title in vector_dbs:
|
123 |
-
retriever = vector_dbs[doc_title].as_retriever(search_kwargs={"k": 3})
|
124 |
-
else:
|
125 |
-
combined_docs = []
|
126 |
-
for db in vector_dbs.values():
|
127 |
-
combined_docs.extend(db.get_relevant_documents(question))
|
128 |
-
if not combined_docs:
|
129 |
-
return "No relevant information found across uploaded documents."
|
130 |
-
temp_db = FAISS.from_documents(combined_docs, embedding_model)
|
131 |
-
retriever = temp_db.as_retriever(search_kwargs={"k": 3})
|
132 |
-
|
133 |
-
relevant_docs = retriever.get_relevant_documents(question)
|
134 |
-
context = "\n\n".join([doc.page_content for doc in relevant_docs])
|
135 |
-
answer = qa_pipeline(question=question, context=context)["answer"]
|
136 |
-
return answer.strip()
|
137 |
-
except Exception as e:
|
138 |
-
st.error(f"Error during question answering: {e}")
|
139 |
-
return "An error occurred while trying to answer the question."
|
140 |
-
|
141 |
-
# ---------------------------------------------
|
142 |
-
# Flashcard Generation
|
143 |
-
# ---------------------------------------------
|
144 |
-
def generate_flashcards(text):
|
145 |
-
flashcards = []
|
146 |
-
seen_terms = set()
|
147 |
-
sentences = nltk.sent_tokenize(text)
|
148 |
-
|
149 |
-
for i, sent in enumerate(sentences):
|
150 |
-
words = nltk.word_tokenize(sent)
|
151 |
-
tagged_words = nltk.pos_tag(words)
|
152 |
-
|
153 |
-
potential_terms = [word for word, tag in tagged_words if tag.startswith('NN') or tag.startswith('NP')]
|
154 |
-
|
155 |
-
for term in potential_terms:
|
156 |
-
if term in seen_terms:
|
157 |
-
continue
|
158 |
-
|
159 |
-
defining_patterns = [r"\b" + re.escape(term) + r"\b\s+is\s+(?:a|an|the)\s+(.+?)(?:\.|,|\n|$)",
|
160 |
-
r"\b" + re.escape(term) + r"\b\s+refers\s+to\s+(.+?)(?:\.|,|\n|$)",
|
161 |
-
r"\b" + re.escape(term) + r"\b\s+means\s+(.+?)(?:\.|,|\n|$)",
|
162 |
-
r"\b" + re.escape(term) + r"\b,\s+defined\s+as\s+(.+?)(?:\.|,|\n|$)",
|
163 |
-
r"\b" + re.escape(term) + r"\b:\s+(.+?)(?:\.|,|\n|$)"]
|
164 |
-
|
165 |
-
potential_definitions = []
|
166 |
-
for pattern in defining_patterns:
|
167 |
-
match = re.search(pattern, sent, re.IGNORECASE)
|
168 |
-
if match and len(match.groups()) >= 1:
|
169 |
-
potential_definitions.append(match.group(1).strip())
|
170 |
-
|
171 |
-
for definition in potential_definitions:
|
172 |
-
if 2 <= len(definition.split()) <= 30:
|
173 |
-
flashcards.append({"term": term, "definition": definition})
|
174 |
-
seen_terms.add(term)
|
175 |
-
break
|
176 |
-
|
177 |
-
if term not in seen_terms and i > 0:
|
178 |
-
prev_sent = sentences[i-1]
|
179 |
-
defining_patterns_prev = [r"The\s+\b" + re.escape(term) + r"\b\s+is\s+(.+?)(?:\.|,|\n|$)",
|
180 |
-
r"This\s+\b" + re.escape(term) + r"\b\s+refers\s+to\s+(.+?)(?:\.|,|\n|$)",
|
181 |
-
r"It\s+means\s+the\s+\b" + re.escape(term) + r"\b\s+(.+?)(?:\.|,|\n|$)"]
|
182 |
-
for pattern in defining_patterns_prev:
|
183 |
-
match = re.search(pattern, prev_sent, re.IGNORECASE)
|
184 |
-
if match and term in sent and len(match.groups()) >= 1:
|
185 |
-
definition = match.group(1).strip()
|
186 |
-
if 2 <= len(definition.split()) <= 30:
|
187 |
-
flashcards.append({"term": term, "definition": definition})
|
188 |
-
seen_terms.add(term)
|
189 |
-
break
|
190 |
-
|
191 |
-
return flashcards
|
192 |
-
|
193 |
-
# ---------------------------------------------
|
194 |
-
# Text to Speech
|
195 |
-
# ---------------------------------------------
|
196 |
-
def read_aloud(text):
|
197 |
-
try:
|
198 |
-
tts = gTTS(text)
|
199 |
-
audio_path = os.path.join(tempfile.gettempdir(), "summary.mp3")
|
200 |
-
tts.save(audio_path)
|
201 |
-
return audio_path
|
202 |
-
except Exception as e:
|
203 |
-
st.error(f"Error during text-to-speech: {e}")
|
204 |
-
return None
|
205 |
|
206 |
-
#
|
207 |
-
#
|
208 |
-
#
|
209 |
-
|
210 |
-
|
211 |
-
|
212 |
-
|
213 |
-
|
214 |
-
|
215 |
-
|
216 |
-
|
217 |
-
|
218 |
-
|
219 |
-
|
220 |
-
|
221 |
-
|
222 |
-
|
223 |
-
|
224 |
-
|
225 |
-
|
226 |
-
|
227 |
-
|
228 |
-
|
229 |
-
|
230 |
-
|
231 |
-
|
232 |
-
|
233 |
-
|
234 |
-
|
235 |
-
|
236 |
-
|
237 |
-
|
238 |
-
|
239 |
-
|
240 |
-
|
241 |
-
|
242 |
-
|
243 |
-
|
244 |
-
|
245 |
-
|
246 |
-
|
247 |
-
|
248 |
-
|
249 |
-
|
250 |
-
|
251 |
-
|
252 |
-
|
253 |
-
|
254 |
-
|
255 |
-
|
256 |
-
|
257 |
-
|
258 |
-
|
259 |
-
|
260 |
-
|
261 |
-
|
262 |
-
|
263 |
-
|
264 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
265 |
|
266 |
# ---------------------------------------------
|
267 |
# Streamlit Interface with Tabs
|
@@ -305,71 +299,71 @@ with tab2:
|
|
305 |
else:
|
306 |
st.info("Please upload and extract a file in the 'Upload & Extract' tab first.")
|
307 |
|
308 |
-
with tab3:
|
309 |
-
|
310 |
-
|
311 |
-
|
312 |
-
|
313 |
-
|
314 |
-
|
315 |
-
|
316 |
-
|
317 |
-
|
318 |
-
|
319 |
-
|
320 |
-
|
321 |
-
|
322 |
-
|
323 |
-
|
324 |
-
|
325 |
-
|
326 |
-
|
327 |
-
|
328 |
-
with tab4:
|
329 |
-
|
330 |
-
|
331 |
-
|
332 |
-
|
333 |
-
|
334 |
-
|
335 |
-
|
336 |
-
|
337 |
-
|
338 |
-
|
339 |
-
|
340 |
-
|
341 |
-
|
342 |
-
|
343 |
-
|
344 |
-
|
345 |
-
|
346 |
-
|
347 |
-
|
348 |
-
|
349 |
-
|
350 |
-
with tab5:
|
351 |
-
|
352 |
-
|
353 |
-
|
354 |
-
|
355 |
-
|
356 |
-
|
357 |
-
|
358 |
-
|
359 |
-
|
360 |
-
|
361 |
-
|
362 |
-
|
363 |
-
|
364 |
-
|
365 |
-
|
366 |
-
|
367 |
-
|
368 |
-
|
369 |
-
|
370 |
-
|
371 |
-
|
372 |
-
|
373 |
-
|
374 |
-
|
375 |
-
|
|
|
21 |
# nltk.download('averaged_perceptron_tagger', download_dir=nltk_data_dir, quiet=True)
|
22 |
# nltk.data.path.append(nltk_data_dir)
|
23 |
|
24 |
+
# nltk_data_dir = "/tmp/nltk_data"
|
25 |
+
# os.makedirs(nltk_data_dir, exist_ok=True)
|
26 |
+
# nltk.data.path.append(nltk_data_dir)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
27 |
|
28 |
+
# # nltk.download('punkt', download_dir=nltk_data_dir)
|
29 |
+
# # nltk.download('averaged_perceptron_tagger', download_dir=nltk_data_dir)
|
30 |
+
# nltk.download('averaged_perceptron_tagger', download_dir=nltk_data_dir, quiet=True)
|
31 |
+
# nltk.download('punkt', download_dir=nltk_data_dir, quiet=True)
|
32 |
+
|
33 |
+
|
34 |
+
|
35 |
+
# # Load models
|
36 |
+
|
37 |
+
# summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
|
38 |
+
# qa_pipeline = pipeline("question-answering", model="distilbert-base-cased-distilled-squad", device=torch.device("cpu"))
|
39 |
+
# embedding_model = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
|
40 |
+
# vector_dbs = {} # Dictionary to store multiple vector databases, keyed by document title
|
41 |
+
# extracted_texts = {} # Dictionary to store extracted text, keyed by document title
|
42 |
+
# current_doc_title = None
|
43 |
+
|
44 |
+
# # ---------------------------------------------
|
45 |
+
# # Extract text from PDF or Image
|
46 |
+
# # ---------------------------------------------
|
47 |
+
# def extract_text(uploaded_file):
|
48 |
+
# global current_doc_title
|
49 |
+
# current_doc_title = uploaded_file.name
|
50 |
+
# suffix = uploaded_file.name.lower()
|
51 |
+
# with tempfile.NamedTemporaryFile(delete=False) as tmp:
|
52 |
+
# tmp.write(uploaded_file.read())
|
53 |
+
# path = tmp.name
|
54 |
+
|
55 |
+
# text = ""
|
56 |
+
# if suffix.endswith(".pdf"):
|
57 |
+
# with pdfplumber.open(path) as pdf:
|
58 |
+
# for page in pdf.pages:
|
59 |
+
# page_text = page.extract_text()
|
60 |
+
# if page_text:
|
61 |
+
# text += page_text + "\n"
|
62 |
+
# else:
|
63 |
+
# try:
|
64 |
+
# text = pytesseract.image_to_string(Image.open(path))
|
65 |
+
# except Exception as e:
|
66 |
+
# st.error(f"Error during OCR for {uploaded_file.name}: {e}")
|
67 |
+
# text = ""
|
68 |
+
|
69 |
+
# os.remove(path)
|
70 |
+
# return text.strip()
|
71 |
+
|
72 |
+
# # ---------------------------------------------
|
73 |
+
# # Store Embeddings in FAISS
|
74 |
+
# # ---------------------------------------------
|
75 |
+
# def store_vector(text):
|
76 |
+
# global vector_dbs, current_doc_title
|
77 |
+
# text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
|
78 |
+
# docs = text_splitter.create_documents([text])
|
79 |
+
# for doc in docs:
|
80 |
+
# doc.metadata = {"title": current_doc_title} # Add document title as metadata
|
81 |
+
|
82 |
+
# if current_doc_title in vector_dbs:
|
83 |
+
# vector_dbs[current_doc_title].add_documents(docs) # Append to existing DB
|
84 |
+
# else:
|
85 |
+
# vector_dbs[current_doc_title] = FAISS.from_documents(docs, embedding_model)
|
86 |
+
|
87 |
+
# # ---------------------------------------------
|
88 |
+
# # Summarize Text
|
89 |
+
# # ---------------------------------------------
|
90 |
+
# def summarize(text):
|
91 |
+
# if len(text.split()) < 100:
|
92 |
+
# return "Text too short to summarize."
|
93 |
+
# chunks = [text[i : i + 1024] for i in range(0, len(text), 1024)]
|
94 |
+
# summaries = []
|
95 |
+
# for chunk in chunks:
|
96 |
+
# max_len = int(len(chunk.split()) * 0.6)
|
97 |
+
# max_len = max(30, min(max_len, 150))
|
98 |
+
# try:
|
99 |
+
# summary = summarizer(chunk, max_length=max_len, min_length=20)[0]["summary_text"]
|
100 |
+
# summaries.append(summary)
|
101 |
+
# except Exception as e:
|
102 |
+
# st.error(f"Error during summarization: {e}")
|
103 |
+
# return "An error occurred during summarization."
|
104 |
+
# return " ".join(summaries)
|
105 |
+
|
106 |
+
# # ---------------------------------------------
|
107 |
+
# # Question Answering
|
108 |
+
# # ---------------------------------------------
|
109 |
+
# def topic_search(question, doc_title=None):
|
110 |
+
# global vector_dbs
|
111 |
+
# if not vector_dbs:
|
112 |
+
# st.warning("Please upload and process a file first in the 'Upload & Extract' tab.")
|
113 |
+
# return ""
|
114 |
+
|
115 |
+
# try:
|
116 |
+
# if doc_title and doc_title in vector_dbs:
|
117 |
+
# retriever = vector_dbs[doc_title].as_retriever(search_kwargs={"k": 3})
|
118 |
+
# else:
|
119 |
+
# combined_docs = []
|
120 |
+
# for db in vector_dbs.values():
|
121 |
+
# combined_docs.extend(db.get_relevant_documents(question))
|
122 |
+
# if not combined_docs:
|
123 |
+
# return "No relevant information found across uploaded documents."
|
124 |
+
# temp_db = FAISS.from_documents(combined_docs, embedding_model)
|
125 |
+
# retriever = temp_db.as_retriever(search_kwargs={"k": 3})
|
126 |
+
|
127 |
+
# relevant_docs = retriever.get_relevant_documents(question)
|
128 |
+
# context = "\n\n".join([doc.page_content for doc in relevant_docs])
|
129 |
+
# answer = qa_pipeline(question=question, context=context)["answer"]
|
130 |
+
# return answer.strip()
|
131 |
+
# except Exception as e:
|
132 |
+
# st.error(f"Error during question answering: {e}")
|
133 |
+
# return "An error occurred while trying to answer the question."
|
134 |
+
|
135 |
+
# # ---------------------------------------------
|
136 |
+
# # Flashcard Generation
|
137 |
+
# # ---------------------------------------------
|
138 |
+
# def generate_flashcards(text):
|
139 |
+
# flashcards = []
|
140 |
+
# seen_terms = set()
|
141 |
+
# sentences = nltk.sent_tokenize(text)
|
142 |
+
|
143 |
+
# for i, sent in enumerate(sentences):
|
144 |
+
# words = nltk.word_tokenize(sent)
|
145 |
+
# tagged_words = nltk.pos_tag(words)
|
146 |
+
|
147 |
+
# potential_terms = [word for word, tag in tagged_words if tag.startswith('NN') or tag.startswith('NP')]
|
148 |
+
|
149 |
+
# for term in potential_terms:
|
150 |
+
# if term in seen_terms:
|
151 |
+
# continue
|
152 |
+
|
153 |
+
# defining_patterns = [r"\b" + re.escape(term) + r"\b\s+is\s+(?:a|an|the)\s+(.+?)(?:\.|,|\n|$)",
|
154 |
+
# r"\b" + re.escape(term) + r"\b\s+refers\s+to\s+(.+?)(?:\.|,|\n|$)",
|
155 |
+
# r"\b" + re.escape(term) + r"\b\s+means\s+(.+?)(?:\.|,|\n|$)",
|
156 |
+
# r"\b" + re.escape(term) + r"\b,\s+defined\s+as\s+(.+?)(?:\.|,|\n|$)",
|
157 |
+
# r"\b" + re.escape(term) + r"\b:\s+(.+?)(?:\.|,|\n|$)"]
|
158 |
+
|
159 |
+
# potential_definitions = []
|
160 |
+
# for pattern in defining_patterns:
|
161 |
+
# match = re.search(pattern, sent, re.IGNORECASE)
|
162 |
+
# if match and len(match.groups()) >= 1:
|
163 |
+
# potential_definitions.append(match.group(1).strip())
|
164 |
+
|
165 |
+
# for definition in potential_definitions:
|
166 |
+
# if 2 <= len(definition.split()) <= 30:
|
167 |
+
# flashcards.append({"term": term, "definition": definition})
|
168 |
+
# seen_terms.add(term)
|
169 |
+
# break
|
170 |
+
|
171 |
+
# if term not in seen_terms and i > 0:
|
172 |
+
# prev_sent = sentences[i-1]
|
173 |
+
# defining_patterns_prev = [r"The\s+\b" + re.escape(term) + r"\b\s+is\s+(.+?)(?:\.|,|\n|$)",
|
174 |
+
# r"This\s+\b" + re.escape(term) + r"\b\s+refers\s+to\s+(.+?)(?:\.|,|\n|$)",
|
175 |
+
# r"It\s+means\s+the\s+\b" + re.escape(term) + r"\b\s+(.+?)(?:\.|,|\n|$)"]
|
176 |
+
# for pattern in defining_patterns_prev:
|
177 |
+
# match = re.search(pattern, prev_sent, re.IGNORECASE)
|
178 |
+
# if match and term in sent and len(match.groups()) >= 1:
|
179 |
+
# definition = match.group(1).strip()
|
180 |
+
# if 2 <= len(definition.split()) <= 30:
|
181 |
+
# flashcards.append({"term": term, "definition": definition})
|
182 |
+
# seen_terms.add(term)
|
183 |
+
# break
|
184 |
+
|
185 |
+
# return flashcards
|
186 |
+
|
187 |
+
# # ---------------------------------------------
|
188 |
+
# # Text to Speech
|
189 |
+
# # ---------------------------------------------
|
190 |
+
# def read_aloud(text):
|
191 |
+
# try:
|
192 |
+
# tts = gTTS(text)
|
193 |
+
# audio_path = os.path.join(tempfile.gettempdir(), "summary.mp3")
|
194 |
+
# tts.save(audio_path)
|
195 |
+
# return audio_path
|
196 |
+
# except Exception as e:
|
197 |
+
# st.error(f"Error during text-to-speech: {e}")
|
198 |
+
# return None
|
199 |
+
|
200 |
+
# # ---------------------------------------------
|
201 |
+
# # Quiz Generation and Handling
|
202 |
+
# # ---------------------------------------------
|
203 |
+
# def generate_quiz_questions(text, num_questions=5):
|
204 |
+
# flashcards = generate_flashcards(text) # Reuse flashcard logic for potential terms/definitions
|
205 |
+
# if not flashcards:
|
206 |
+
# return []
|
207 |
+
|
208 |
+
# questions = []
|
209 |
+
# used_indices = set()
|
210 |
+
# num_available = len(flashcards)
|
211 |
+
|
212 |
+
# while len(questions) < num_questions and len(used_indices) < num_available:
|
213 |
+
# index = random.randint(0, num_available - 1)
|
214 |
+
# if index in used_indices:
|
215 |
+
# continue
|
216 |
+
# used_indices.add(index)
|
217 |
+
# card = flashcards[index]
|
218 |
+
# correct_answer = card['term']
|
219 |
+
# definition = card['definition']
|
220 |
+
|
221 |
+
# # Generate incorrect answers (very basic for now)
|
222 |
+
# incorrect_options = random.sample([c['term'] for i, c in enumerate(flashcards) if i != index], 3)
|
223 |
+
# options = [correct_answer] + incorrect_options
|
224 |
+
# random.shuffle(options)
|
225 |
+
|
226 |
+
# questions.append({
|
227 |
+
# "question": f"What is the term for: {definition}",
|
228 |
+
# "options": options,
|
229 |
+
# "correct_answer": correct_answer,
|
230 |
+
# "user_answer": None # To store user's choice
|
231 |
+
# })
|
232 |
+
|
233 |
+
# return questions
|
234 |
+
|
235 |
+
# def display_quiz(questions):
|
236 |
+
# st.session_state.quiz_questions = questions
|
237 |
+
# st.session_state.user_answers = {}
|
238 |
+
# st.session_state.quiz_submitted = False
|
239 |
+
# for i, q in enumerate(st.session_state.quiz_questions):
|
240 |
+
# st.subheader(f"Question {i + 1}:")
|
241 |
+
# st.write(q["question"])
|
242 |
+
# st.session_state.user_answers[i] = st.radio(f"Answer for Question {i + 1}", q["options"])
|
243 |
+
# st.button("Submit Quiz", on_click=submit_quiz)
|
244 |
+
|
245 |
+
# def submit_quiz():
|
246 |
+
# st.session_state.quiz_submitted = True
|
247 |
+
|
248 |
+
# def grade_quiz():
|
249 |
+
# if st.session_state.quiz_submitted:
|
250 |
+
# score = 0
|
251 |
+
# for i, q in enumerate(st.session_state.quiz_questions):
|
252 |
+
# user_answer = st.session_state.user_answers.get(i)
|
253 |
+
# if user_answer == q["correct_answer"]:
|
254 |
+
# score += 1
|
255 |
+
# st.success(f"Question {i + 1}: Correct!")
|
256 |
+
# else:
|
257 |
+
# st.error(f"Question {i + 1}: Incorrect. Correct answer was: {q['correct_answer']}")
|
258 |
+
# st.write(f"## Your Score: {score} / {len(st.session_state.quiz_questions)}")
|
259 |
|
260 |
# ---------------------------------------------
|
261 |
# Streamlit Interface with Tabs
|
|
|
299 |
else:
|
300 |
st.info("Please upload and extract a file in the 'Upload & Extract' tab first.")
|
301 |
|
302 |
+
# with tab3:
|
303 |
+
# st.header("❓ Question Answering")
|
304 |
+
# doc_titles = list(vector_dbs.keys())
|
305 |
+
# if doc_titles:
|
306 |
+
# doc_title = st.selectbox("Search within document:", ["All Documents"] + doc_titles)
|
307 |
+
# question = st.text_input("Ask a question about the content:")
|
308 |
+
# if question:
|
309 |
+
# with st.spinner("Searching for answer..."):
|
310 |
+
# if doc_title == "All Documents":
|
311 |
+
# answer = topic_search(question)
|
312 |
+
# else:
|
313 |
+
# answer = topic_search(question, doc_title=doc_title)
|
314 |
+
# if answer:
|
315 |
+
# st.subheader("Answer:")
|
316 |
+
# st.write(answer)
|
317 |
+
# else:
|
318 |
+
# st.warning("Could not find an answer in the selected document(s).")
|
319 |
+
# else:
|
320 |
+
# st.info("Please upload and extract a file in the 'Upload & Extract' tab first.")
|
321 |
+
|
322 |
+
# with tab4:
|
323 |
+
# st.header("🧠 Interactive Learning: Flashcards")
|
324 |
+
# doc_titles = list(extracted_texts.keys())
|
325 |
+
# if doc_titles:
|
326 |
+
# selected_doc_title_flashcard = st.selectbox("Generate flashcards from document:", doc_titles)
|
327 |
+
# if st.button("Generate Flashcards"):
|
328 |
+
# if selected_doc_title_flashcard in extracted_texts:
|
329 |
+
# with st.spinner(f"Generating flashcards from {selected_doc_title_flashcard}..."):
|
330 |
+
# flashcards = generate_flashcards(extracted_texts[selected_doc_title_flashcard])
|
331 |
+
# if flashcards:
|
332 |
+
# st.subheader("Flashcards")
|
333 |
+
# for i, card in enumerate(flashcards):
|
334 |
+
# with st.expander(f"Card {i+1}"):
|
335 |
+
# st.markdown(f"*Term:* {card['term']}")
|
336 |
+
# st.markdown(f"*Definition:* {card['definition']}")
|
337 |
+
# else:
|
338 |
+
# st.info("No flashcards could be generated from this document using the current method.")
|
339 |
+
# else:
|
340 |
+
# st.warning(f"Original text for {selected_doc_title_flashcard} not found. Please re-upload.")
|
341 |
+
# else:
|
342 |
+
# st.info("Please upload and extract a file in the 'Upload & Extract' tab first.")
|
343 |
+
|
344 |
+
# with tab5:
|
345 |
+
# st.header("📝 Quiz Yourself!")
|
346 |
+
# doc_titles = list(extracted_texts.keys())
|
347 |
+
# if doc_titles:
|
348 |
+
# selected_doc_title_quiz = st.selectbox("Generate quiz from document:", doc_titles)
|
349 |
+
# if selected_doc_title_quiz in extracted_texts:
|
350 |
+
# text_for_quiz =extracted_texts[selected_doc_title_quiz]
|
351 |
+
# if "quiz_questions" not in st.session_state:
|
352 |
+
# st.session_state.quiz_questions = generate_quiz_questions(text_for_quiz)
|
353 |
+
|
354 |
+
# if st.session_state.quiz_questions:
|
355 |
+
# display_quiz(st.session_state.quiz_questions)
|
356 |
+
# if st.session_state.quiz_submitted:
|
357 |
+
# grade_quiz()
|
358 |
+
|
359 |
+
# if st.button("Refresh Questions"):
|
360 |
+
# st.session_state.quiz_questions = generate_quiz_questions(text_for_quiz)
|
361 |
+
# st.session_state.quiz_submitted = False
|
362 |
+
# st.session_state.user_answers = {}
|
363 |
+
# st.rerun() # Force a re-render to show new questions
|
364 |
+
# else:
|
365 |
+
# st.info("Could not generate quiz questions from the current document.")
|
366 |
+
# else:
|
367 |
+
# st.warning(f"Original text for {selected_doc_title_quiz} not found. Please re-upload.")
|
368 |
+
# else:
|
369 |
+
# st.info("Please upload and extract a file in the 'Upload & Extract' tab first.")
|