Spaces:
Running
Running
adibak
commited on
Commit
·
dade368
1
Parent(s):
b4b5aad
move file input to sidebar and set page range max to pdf length
Browse files- app.py +43 -13
- helpers/file_manager.py +9 -4
app.py
CHANGED
|
@@ -221,12 +221,43 @@ with st.sidebar:
|
|
| 221 |
),
|
| 222 |
value='2024-05-01-preview',
|
| 223 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 224 |
|
| 225 |
-
page_range_slider = st.slider(label=('4: Specify a page range to examine:\n\n'
|
| 226 |
-
'(min=1, max=50)'),
|
| 227 |
-
min_value=1, max_value=50,
|
| 228 |
-
value=(1, 50))
|
| 229 |
-
st.session_state['page_range'] = page_range_slider
|
| 230 |
|
| 231 |
|
| 232 |
def build_ui():
|
|
@@ -283,18 +314,17 @@ def set_up_chat_ui():
|
|
| 283 |
if prompt := st.chat_input(
|
| 284 |
placeholder=APP_TEXT['chat_placeholder'],
|
| 285 |
max_chars=GlobalConfig.LLM_MODEL_MAX_INPUT_LENGTH,
|
| 286 |
-
accept_file=
|
| 287 |
file_type=['pdf', ],
|
| 288 |
):
|
| 289 |
-
|
| 290 |
-
|
| 291 |
-
|
| 292 |
-
|
| 293 |
-
page_range = st.session_state.get('page_range', (1, 50)) # fallback default
|
| 294 |
st.session_state[ADDITIONAL_INFO] = filem.get_pdf_contents(
|
| 295 |
-
|
|
|
|
| 296 |
)
|
| 297 |
-
print(f'{prompt["files"]=}')
|
| 298 |
|
| 299 |
provider, llm_name = llm_helper.get_provider_model(
|
| 300 |
llm_provider_to_use,
|
|
|
|
| 221 |
),
|
| 222 |
value='2024-05-01-preview',
|
| 223 |
)
|
| 224 |
+
|
| 225 |
+
from pypdf import PdfReader
|
| 226 |
+
|
| 227 |
+
uploaded_pdf = st.file_uploader("1: Upload a PDF file", type=["pdf"])
|
| 228 |
+
|
| 229 |
+
# Detect file change: update session state
|
| 230 |
+
if uploaded_pdf:
|
| 231 |
+
# Unique hash for reset logic
|
| 232 |
+
new_file_hash = hash(uploaded_pdf.getvalue())
|
| 233 |
+
|
| 234 |
+
# If file is newly uploaded or changed
|
| 235 |
+
if st.session_state.get("current_pdf_hash") != new_file_hash:
|
| 236 |
+
reader = PdfReader(uploaded_pdf)
|
| 237 |
+
total_pages = len(reader.pages)
|
| 238 |
+
|
| 239 |
+
st.session_state["pdf_page_count"] = total_pages
|
| 240 |
+
st.session_state["current_pdf_hash"] = new_file_hash
|
| 241 |
+
|
| 242 |
+
# Force slider reset
|
| 243 |
+
st.session_state.pop("page_range", None)
|
| 244 |
+
|
| 245 |
+
# Set default page count if no file uploaded
|
| 246 |
+
page_count = st.session_state.get("pdf_page_count", 50)
|
| 247 |
+
max_slider = min(50, page_count)
|
| 248 |
+
|
| 249 |
+
# Show slider only after file upload
|
| 250 |
+
if "pdf_page_count" in st.session_state:
|
| 251 |
+
page_range_slider = st.slider(
|
| 252 |
+
label="2: Specify a page range to examine:",
|
| 253 |
+
min_value=1,
|
| 254 |
+
max_value=max_slider,
|
| 255 |
+
value=(1, max_slider),
|
| 256 |
+
key="page_range" # persistent + resettable
|
| 257 |
+
)
|
| 258 |
+
else:
|
| 259 |
+
st.info("📄 Upload a PDF to specify a page range.")
|
| 260 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 261 |
|
| 262 |
|
| 263 |
def build_ui():
|
|
|
|
| 314 |
if prompt := st.chat_input(
|
| 315 |
placeholder=APP_TEXT['chat_placeholder'],
|
| 316 |
max_chars=GlobalConfig.LLM_MODEL_MAX_INPUT_LENGTH,
|
| 317 |
+
accept_file=False,
|
| 318 |
file_type=['pdf', ],
|
| 319 |
):
|
| 320 |
+
logger.info(f"type {type(prompt)}")
|
| 321 |
+
prompt_text = prompt
|
| 322 |
+
|
| 323 |
+
if uploaded_pdf and "page_range" in st.session_state:
|
|
|
|
| 324 |
st.session_state[ADDITIONAL_INFO] = filem.get_pdf_contents(
|
| 325 |
+
uploaded_pdf,
|
| 326 |
+
st.session_state["page_range"]
|
| 327 |
)
|
|
|
|
| 328 |
|
| 329 |
provider, llm_name = llm_helper.get_provider_model(
|
| 330 |
llm_provider_to_use,
|
helpers/file_manager.py
CHANGED
|
@@ -26,19 +26,24 @@ def get_pdf_contents(
|
|
| 26 |
Extract the text contents from a PDF file.
|
| 27 |
|
| 28 |
:param pdf_file: The uploaded PDF file.
|
|
|
|
| 29 |
:param max_pages: The max no. of pages to extract contents from.
|
| 30 |
:return: The contents.
|
| 31 |
"""
|
| 32 |
|
| 33 |
reader = PdfReader(pdf_file)
|
| 34 |
|
|
|
|
| 35 |
total_pages = len(reader.pages)
|
| 36 |
-
n_pages = min(max_pages, total_pages)
|
| 37 |
|
| 38 |
-
|
| 39 |
-
start =
|
| 40 |
-
|
|
|
|
| 41 |
|
|
|
|
|
|
|
| 42 |
text = ''
|
| 43 |
for page_num in range(start - 1, end):
|
| 44 |
page = reader.pages[page_num]
|
|
|
|
| 26 |
Extract the text contents from a PDF file.
|
| 27 |
|
| 28 |
:param pdf_file: The uploaded PDF file.
|
| 29 |
+
:param page_range: The range of pages to extract contents from.
|
| 30 |
:param max_pages: The max no. of pages to extract contents from.
|
| 31 |
:return: The contents.
|
| 32 |
"""
|
| 33 |
|
| 34 |
reader = PdfReader(pdf_file)
|
| 35 |
|
| 36 |
+
# get number of pages
|
| 37 |
total_pages = len(reader.pages)
|
| 38 |
+
n_pages = min(max_pages, total_pages) # set n_pages to min of 50 or the pdf length if the pdf is shorter than 50 pages
|
| 39 |
|
| 40 |
+
# ensure validity
|
| 41 |
+
start, end = page_range # set start and end per the range (user-specified values)
|
| 42 |
+
start = max(1, start) # set start to max of 1, or user-spec start
|
| 43 |
+
end = min(n_pages, end) # set end to min of n_pages, or user-spec end
|
| 44 |
|
| 45 |
+
logger.info(f"start={start}")
|
| 46 |
+
logger.info(f"end={end}")
|
| 47 |
text = ''
|
| 48 |
for page_num in range(start - 1, end):
|
| 49 |
page = reader.pages[page_num]
|