Spaces:
Sleeping
Sleeping
elia-waefler
commited on
Commit
·
f91d70b
1
Parent(s):
7d56215
added mock functionality
Browse files
app.py
CHANGED
@@ -118,7 +118,6 @@ def handle_userinput(user_question):
|
|
118 |
st.write(f"Source Document: {message.source}", unsafe_allow_html=True)
|
119 |
|
120 |
|
121 |
-
|
122 |
def safe_vec_store():
|
123 |
# USE VECTARA INSTEAD
|
124 |
os.makedirs('vectorstore', exist_ok=True)
|
@@ -135,40 +134,42 @@ def main():
|
|
135 |
load_dotenv()
|
136 |
st.set_page_config(page_title="Doc Verify RAG", page_icon=":hospital:")
|
137 |
st.write(css, unsafe_allow_html=True)
|
138 |
-
st.session_state
|
|
|
139 |
st.subheader("Your documents")
|
140 |
-
|
141 |
-
|
142 |
-
|
|
|
|
|
143 |
if st.button("Process"):
|
144 |
with st.spinner("Processing"):
|
145 |
if st.session_state.classify:
|
146 |
# THE CLASSIFICATION APP
|
147 |
-
|
148 |
-
|
149 |
-
|
150 |
-
|
151 |
-
|
152 |
-
|
153 |
-
|
154 |
-
|
155 |
-
|
156 |
-
|
157 |
-
|
158 |
-
|
159 |
-
|
160 |
-
|
161 |
-
|
162 |
-
|
163 |
-
|
164 |
-
|
165 |
-
|
166 |
-
|
|
|
|
|
|
|
167 |
st.success("data loaded")
|
168 |
-
if st.session_state.classify:
|
169 |
-
# THE CLASSIFICATION APP
|
170 |
-
classification_result = extract_metadata(plain_text_doc)
|
171 |
-
st.write(classification_result)
|
172 |
|
173 |
|
174 |
if "conversation" not in st.session_state:
|
@@ -180,7 +181,6 @@ def main():
|
|
180 |
user_question = st.text_input("Ask a question about your documents:")
|
181 |
if user_question:
|
182 |
handle_userinput(user_question)
|
183 |
-
|
184 |
with st.sidebar:
|
185 |
|
186 |
st.subheader("Classification Instrucitons")
|
@@ -188,8 +188,9 @@ def main():
|
|
188 |
filenames = [file.name for file in classifier_docs if file is not None]
|
189 |
|
190 |
if st.button("Process Classification"):
|
|
|
191 |
with st.spinner("Processing"):
|
192 |
-
st.
|
193 |
time.sleep(3)
|
194 |
|
195 |
|
|
|
118 |
st.write(f"Source Document: {message.source}", unsafe_allow_html=True)
|
119 |
|
120 |
|
|
|
121 |
def safe_vec_store():
|
122 |
# USE VECTARA INSTEAD
|
123 |
os.makedirs('vectorstore', exist_ok=True)
|
|
|
134 |
load_dotenv()
|
135 |
st.set_page_config(page_title="Doc Verify RAG", page_icon=":hospital:")
|
136 |
st.write(css, unsafe_allow_html=True)
|
137 |
+
if not "classify" in st.session_state:
|
138 |
+
st.session_state.classify = False
|
139 |
st.subheader("Your documents")
|
140 |
+
if st.session_state.classify:
|
141 |
+
pdf_doc = st.file_uploader("Upload your PDFs here and click on 'Process'", accept_multiple_files=False)
|
142 |
+
else:
|
143 |
+
pdf_docs = st.file_uploader("Upload your PDFs here and click on 'Process'", accept_multiple_files=True)
|
144 |
+
filenames = [file.name for file in pdf_docs if file is not None]
|
145 |
if st.button("Process"):
|
146 |
with st.spinner("Processing"):
|
147 |
if st.session_state.classify:
|
148 |
# THE CLASSIFICATION APP
|
149 |
+
st.write("Classifying")
|
150 |
+
plain_text_doc = ingest(pdf_doc.name)
|
151 |
+
classification_result = extract_metadata(plain_text_doc)
|
152 |
+
st.write(classification_result)
|
153 |
+
else:
|
154 |
+
# NORMAL RAG
|
155 |
+
loaded_vec_store = None
|
156 |
+
for filename in filenames:
|
157 |
+
if ".pkl" in filename:
|
158 |
+
file_path = os.path.join('vectorstore', filename)
|
159 |
+
with open(file_path, 'rb') as f:
|
160 |
+
loaded_vec_store = pickle.load(f)
|
161 |
+
raw_text = get_pdf_text(pdf_docs)
|
162 |
+
text_chunks = get_text_chunks(raw_text)
|
163 |
+
vec = get_vectorstore(text_chunks)
|
164 |
+
if loaded_vec_store:
|
165 |
+
vec.merge_from(loaded_vec_store)
|
166 |
+
st.warning("loaded vectorstore")
|
167 |
+
if "vectorstore" in st.session_state:
|
168 |
+
vec.merge_from(st.session_state.vectorstore)
|
169 |
+
st.warning("merged to existing")
|
170 |
+
st.session_state.vectorstore = vec
|
171 |
+
st.session_state.conversation = get_conversation_chain(vec)
|
172 |
st.success("data loaded")
|
|
|
|
|
|
|
|
|
173 |
|
174 |
|
175 |
if "conversation" not in st.session_state:
|
|
|
181 |
user_question = st.text_input("Ask a question about your documents:")
|
182 |
if user_question:
|
183 |
handle_userinput(user_question)
|
|
|
184 |
with st.sidebar:
|
185 |
|
186 |
st.subheader("Classification Instrucitons")
|
|
|
188 |
filenames = [file.name for file in classifier_docs if file is not None]
|
189 |
|
190 |
if st.button("Process Classification"):
|
191 |
+
st.session_state.classify = True
|
192 |
with st.spinner("Processing"):
|
193 |
+
st.warning("set classify")
|
194 |
time.sleep(3)
|
195 |
|
196 |
|