Update app.py

app.py

@@ -6,10 +6,11 @@ import openai
 import pandas as pd
 from IPython.display import Markdown, display
 from langchain.document_loaders import PyPDFLoader
+from langchain.embeddings import OpenAIEmbeddings
 from langchain.indexes import VectorstoreIndexCreator
+from langchain.llms import OpenAI
 from langchain.vectorstores import DocArrayInMemorySearch
-from langchain.embeddings import OpenAIEmbeddings
-
+from uuid import uuid4
 
 css_style = """
 .gradio-container {
@@ -24,6 +25,7 @@ class myClass:
         self.valid_key = False
         self.docs_ready = False
         self.status = "⚠️Waiting for documents and key⚠️"
+        self.uuid = uuid4()
         pass
 
     def check_status(self):
@@ -42,9 +44,11 @@ class myClass:
         assert isinstance(myin, str)
         self.valid_key = True
         self.openai_api_key = myin.strip()
-    
+        self.embedding = OpenAIEmbeddings(openai_api_key=self.openai_api_key)
+        self.llm = OpenAI(openai_api_key=self.openai_api_key)
+
         self.check_status()
-        return self.status
+        return [self.status]
 
     def request_pathname(self, files, data):
         if files is None:
@@ -78,58 +82,70 @@ class myClass:
 
     def get_index(self):
         if self.docs_ready and self.valid_key:
-            # openai = OpenAIEmbeddings(openai_api_key=self.openai_api_key)
-            os.environ["OPENAI_API_KEY"] = self.openai_api_key
-
+            # os.environ["OPENAI_API_KEY"] = self.openai_api_key
 
             # myfile = "Angela Merkel - Wikipedia.pdf"
             # loader = PyPDFLoader(file_path=myfile)
-            loader = PyPDFLoader(file_path=self.dataset["filepath"][0])
+            loaders = [PyPDFLoader(f) for f in self.dataset["filepath"]]
 
             self.index = VectorstoreIndexCreator(
-                vectorstore_cls=DocArrayInMemorySearch
-            ).from_loaders([loader])
-            del os.environ["OPENAI_API_KEY"]
+                vectorstore_cls=DocArrayInMemorySearch, embedding=self.embedding
+            ).from_loaders(loaders=loaders)
+
+            # del os.environ["OPENAI_API_KEY"]
 
         pass
 
     def do_ask(self, question):
         # os.environ["OPENAI_API_KEY"] = self.openai_api_key
-        # openai.api_key = self.openai_api_key    
+        # openai.api_key = self.openai_api_key
+
         if self.status == "✨Ready✨":
-            # openai = OpenAIEmbeddings(openai_api_key=self.openai_api_key)
-            os.environ["OPENAI_API_KEY"] = self.openai_api_key
-            
-            response = self.index.query(question=question)
-            del os.environ["OPENAI_API_KEY"]
+            # os.environ["OPENAI_API_KEY"] = self.openai_api_key
+
+            response = self.index.query(question=question, llm=self.llm)
+            # del os.environ["OPENAI_API_KEY"]
             yield response
         pass
 
 
+def validate_key(myInstance: myClass, openai_api_key):
+    if myInstance is None:
+        myInstance = myClass()
+
+    out = myInstance.validate_key(openai_api_key)
+    return myInstance, *out
+
+
+def request_pathname(myInstance: myClass, files, data):
+    if myInstance is None:
+        myInstance = myClass()
+    out = myInstance.request_pathname(files, data)
+    return myInstance, *out
+
+
+def do_ask(myInstance: myClass, question):
+    out = myInstance.do_ask(question)
+    return myInstance, *out
+
+
 with gr.Blocks(css=css_style) as demo:
+    myInstance = gr.State()
+    openai_api_key = gr.State("")
     docs = gr.State()
     data = gr.State([])
-    openai_api_key = gr.State("")
     index = gr.State()
-    myInstance = gr.State()
-    myInstance = myClass()
 
     gr.Markdown(
         """
     # Document Question and Answer
-    
     *By D8a.ai*
-
     Idea based on https://huggingface.co/spaces/whitead/paper-qa
-    
     Significant advances in langchain have made it possible to simplify the code.
-
     This tool allows you to ask questions of your uploaded text, PDF documents.
-
     It uses OpenAI's GPT models, so you need to enter your API key below. This
     tool is under active development and currently uses a lot of tokens - up to 10,000
     for a single query. This is $0.10-0.20 per query, so please be careful!
-    
     * [langchain](https://github.com/hwchase17/langchain) is the main library this tool utilizes.
     1. Enter API Key ([What is that?](https://platform.openai.com/account/api-keys))
     2. Upload your documents
@@ -169,21 +185,21 @@ with gr.Blocks(css=css_style) as demo:
     answer = gr.Markdown(label="Answer")
 
     openai_api_key.change(
-        myInstance.validate_key, inputs=openai_api_key, outputs=buildb
+        validate_key, inputs=[myInstance, openai_api_key], outputs=[myInstance, buildb]
     )
 
     uploaded_files.change(
-        myInstance.request_pathname,
-        inputs=[uploaded_files, data],
-        outputs=[dataset, buildb],
+        request_pathname,
+        inputs=[myInstance, uploaded_files, data],
+        outputs=[myInstance, dataset, buildb],
     )
 
     ask.click(
-        myInstance.do_ask,
-        inputs=[query],
-        outputs=answer,
+        do_ask,
+        inputs=[myInstance, query],
+        outputs=[myInstance, answer],
     )
 
 
 demo.queue(concurrency_count=20)
-demo.launch(show_error=True)
+demo.launch(show_error=True)