Update app.py
Browse files
app.py
CHANGED
|
@@ -16,7 +16,7 @@ def read_doc(directory):
|
|
| 16 |
documents=file_loader.load()
|
| 17 |
return documents
|
| 18 |
|
| 19 |
-
doc=read_doc('documents/')
|
| 20 |
len(doc)
|
| 21 |
|
| 22 |
## Divide the docs into chunks
|
|
@@ -26,7 +26,7 @@ def chunk_data(docs,chunk_size=800,chunk_overlap=50):
|
|
| 26 |
doc=text_splitter.split_documents(docs)
|
| 27 |
return docs
|
| 28 |
|
| 29 |
-
documents=chunk_data(docs=doc)
|
| 30 |
len(documents)
|
| 31 |
|
| 32 |
embeddings=OpenAIEmbeddings(api_key=os.environ['OPENAI_API_KEY'])
|
|
@@ -36,10 +36,10 @@ vectors=embeddings.embed_query("How are you?")
|
|
| 36 |
len(vectors)
|
| 37 |
|
| 38 |
pinecone.init(
|
| 39 |
-
api_key="
|
| 40 |
environment="gcp-starter"
|
| 41 |
)
|
| 42 |
-
index_name="
|
| 43 |
|
| 44 |
index=Pinecone.from_documents(doc,embeddings,index_name=index_name)
|
| 45 |
|
|
@@ -62,6 +62,6 @@ def retrieve_answers(query):
|
|
| 62 |
return response
|
| 63 |
|
| 64 |
|
| 65 |
-
our_query = "
|
| 66 |
answer = retrieve_answers(our_query)
|
| 67 |
print(answer)
|
|
|
|
| 16 |
documents=file_loader.load()
|
| 17 |
return documents
|
| 18 |
|
| 19 |
+
doc=read_doc('documents/') #PDF directory
|
| 20 |
len(doc)
|
| 21 |
|
| 22 |
## Divide the docs into chunks
|
|
|
|
| 26 |
doc=text_splitter.split_documents(docs)
|
| 27 |
return docs
|
| 28 |
|
| 29 |
+
documents=chunk_data(docs=doc) #Fnct that divides the PDF into chuncks
|
| 30 |
len(documents)
|
| 31 |
|
| 32 |
embeddings=OpenAIEmbeddings(api_key=os.environ['OPENAI_API_KEY'])
|
|
|
|
| 36 |
len(vectors)
|
| 37 |
|
| 38 |
pinecone.init(
|
| 39 |
+
api_key="3cdc872c-aecc-4b11-93d6-b5243930ac3a",
|
| 40 |
environment="gcp-starter"
|
| 41 |
)
|
| 42 |
+
index_name="knowledgebase"
|
| 43 |
|
| 44 |
index=Pinecone.from_documents(doc,embeddings,index_name=index_name)
|
| 45 |
|
|
|
|
| 62 |
return response
|
| 63 |
|
| 64 |
|
| 65 |
+
our_query = "What is my name?"
|
| 66 |
answer = retrieve_answers(our_query)
|
| 67 |
print(answer)
|