giving user ability to upload PDF

app.py

@@ -1,4 +1,5 @@
 import gradio as gr
+from gradio_pdf import PDF
 from qdrant_client import models, QdrantClient
 from sentence_transformers import SentenceTransformer
 from PyPDF2 import PdfReader
@@ -45,7 +46,75 @@ llm = AutoModelForCausalLM.from_pretrained("TheBloke/Llama-2-7B-Chat-GGUF",
 print("LLM loaded........................................")
 print("################################################################")
 
-def get_chunks(text):
+# def get_chunks(text):
+#     text_splitter = RecursiveCharacterTextSplitter(
+#         # seperator = "\n",
+#         chunk_size = 250,
+#         chunk_overlap = 50,
+#         length_function = len,
+#     )
+
+#     chunks = text_splitter.split_text(text)
+#     return chunks
+
+
+# pdf_path = './100 Weird Facts About the Human Body.pdf'
+
+
+# reader = PdfReader(pdf_path)
+# text = ""
+# num_of_pages = len(reader.pages)
+
+# for page in range(num_of_pages):
+#     current_page = reader.pages[page]
+#     text += current_page.extract_text()
+
+
+# chunks = get_chunks(text)
+# print(chunks)
+# print("Chunks are ready.....................................")
+# print("######################################################")
+
+# client = QdrantClient(path = "./db")
+# print("db  created................................................")
+# print("#####################################################################")
+
+# client.recreate_collection(
+#     collection_name="my_facts",
+#     vectors_config=models.VectorParams(
+#         size=encoder.get_sentence_embedding_dimension(),  # Vector size is defined by used model
+#         distance=models.Distance.COSINE,
+#     ),
+# )
+
+# print("Collection created........................................")
+# print("#########################################################")
+
+
+
+# li = []
+# for i in range(len(chunks)):
+#     li.append(i)
+ 
+# dic = zip(li, chunks)
+# dic= dict(dic)
+
+# client.upload_records(
+#     collection_name="my_facts",
+#     records=[
+#         models.Record(
+#             id=idx,
+#             vector=encoder.encode(dic[idx]).tolist(),
+#             payload= {dic[idx][:5] : dic[idx]}
+#         ) for idx in dic.keys()
+#     ],
+# )
+
+# print("Records uploaded........................................")
+# print("###########################################################")
+
+def chat(file, question):
+    def get_chunks(text):
     text_splitter = RecursiveCharacterTextSplitter(
         # seperator = "\n",
         chunk_size = 250,
@@ -57,62 +126,61 @@ def get_chunks(text):
     return chunks
 
 
-pdf_path = './100 Weird Facts About the Human Body.pdf'
+    pdf_path = './100 Weird Facts About the Human Body.pdf'
 
 
-reader = PdfReader(pdf_path)
-text = ""
-num_of_pages = len(reader.pages)
+    reader = PdfReader(pdf_path)
+    text = ""
+    num_of_pages = len(reader.pages)
 
-for page in range(num_of_pages):
-    current_page = reader.pages[page]
-    text += current_page.extract_text()
+    for page in range(num_of_pages):
+        current_page = reader.pages[page]
+        text += current_page.extract_text()
 
 
-chunks = get_chunks(text)
-print(chunks)
-print("Chunks are ready.....................................")
-print("######################################################")
+    chunks = get_chunks(text)
+    # print(chunks)
+# print("Chunks are ready.....................................")
+# print("######################################################")
 
-client = QdrantClient(path = "./db")
-print("db  created................................................")
-print("#####################################################################")
+    client = QdrantClient(path = "./db")
+# print("db  created................................................")
+# print("#####################################################################")
 
-client.recreate_collection(
-    collection_name="my_facts",
-    vectors_config=models.VectorParams(
-        size=encoder.get_sentence_embedding_dimension(),  # Vector size is defined by used model
-        distance=models.Distance.COSINE,
-    ),
-)
+    client.recreate_collection(
+        collection_name="my_facts",
+        vectors_config=models.VectorParams(
+            size=encoder.get_sentence_embedding_dimension(),  # Vector size is defined by used model
+            distance=models.Distance.COSINE,
+        ),
+    )
 
-print("Collection created........................................")
-print("#########################################################")
+# print("Collection created........................................")
+# print("#########################################################")
 
 
 
-li = []
-for i in range(len(chunks)):
-    li.append(i)
+    li = []
+    for i in range(len(chunks)):
+        li.append(i)
  
-dic = zip(li, chunks)
-dic= dict(dic)
-
-client.upload_records(
-    collection_name="my_facts",
-    records=[
-        models.Record(
-            id=idx,
-            vector=encoder.encode(dic[idx]).tolist(),
-            payload= {dic[idx][:5] : dic[idx]}
-        ) for idx in dic.keys()
-    ],
-)
+    dic = zip(li, chunks)
+    dic= dict(dic)
+
+    client.upload_records(
+        collection_name="my_facts",
+        records=[
+            models.Record(
+                id=idx,
+                vector=encoder.encode(dic[idx]).tolist(),
+                payload= {dic[idx][:5] : dic[idx]}
+            ) for idx in dic.keys()
+        ],
+    )
 
-print("Records uploaded........................................")
-print("###########################################################")
+# print("Records uploaded........................................")
+# print("###########################################################")
 
-def chat(question):
 
     hits = client.search(
         collection_name="my_facts",
@@ -148,7 +216,7 @@ def chat(question):
 
 screen = gr.Interface(
     fn = chat,
-    inputs = gr.Textbox(lines = 10, placeholder = "Enter your question here 👉"),
+    inputs = [PDF(label="Upload a PDF", interactive=True), gr.Textbox(lines = 10, placeholder = "Enter your question here 👉")],
     outputs = gr.Textbox(lines = 10, placeholder = "Your answer will be here soon 🚀"),
     title="Q&A with PDF 👩🏻‍💻📓✍🏻💡",
 description="This app facilitates a conversation with PDFs available on https://www.delo.si/assets/media/other/20110728/100%20Weird%20Facts%20About%20the%20Human%20Body.pdf💡",