big update

.gitignore

@@ -0,0 +1,3 @@
+asdf
+__pycache__
+env

app.py

@@ -10,9 +10,15 @@ import pandas as pd
 import langchain
 from langchain import PromptTemplate, LLMChain
 from langchain.chat_models import ChatOpenAI
+from datetime import date
+import numpy as np
+from openai import OpenAI
 
 model_en = SentenceTransformer("intfloat/multilingual-e5-base")
-intents_embedding = create_embedding(intents_sentence_similarity_en, model_en)
+with open('embeddings.npy', 'rb') as f:
+  intents_embedding = np.load(f)
+with open('openai_embeddings.npy', 'rb') as f:
+  openai_intents_embedding = np.load(f)
 
 llm = None
 llm_chain = None
@@ -95,9 +101,10 @@ def check_accuracy(n_samples, threshold):
   print(score, grouped_data)
   return score, grouped_data
 
-def classify_intent(input_text:str, history:str, answer, api_key):
-  # os.environ['OPENAI_API_KEY'] = api_key
-  llm = ChatOpenAI(model='gpt-3.5-turbo', temperature='0.1', openai_api_key=api_key)
+def classify_intent(input_text:str, history:str, answer, model_name, api_key):
+  print(f"predicting with llm... date: {date.today()}")
+  print(f"model name: {model_name}")
+  llm = ChatOpenAI(model=model_name, temperature='0.1')
   prompt = PromptTemplate(template=prompt_template, input_variables=["intents", "INPUT", "chatHistory"])
   llm_chain = LLMChain(prompt=prompt, llm=llm, verbose=False)
 
@@ -119,8 +126,14 @@ with gr.Blocks(title="Intent Classification Demo", theme=theme) as interface:
 
   with gr.Row(equal_height=True):
     with gr.Column():
+      model_name = gr.Dropdown(["gpt-3.5-turbo",
+                            "gpt-3.5-turbo-1106",
+                            "gpt-4",
+                            "gpt-4-1106-preview"],
+                            label="Model name",
+                            info="Select model name for GPT")
       api_key = gr.Textbox(label="OpenAI API Key", info="get it at https://platform.openai.com/account/api-keys",visible=True, lines=1, type="password")
-      n_samples = gr.Slider(1, 10, value=5, step=1, label="N samples", info="Number of samples to be retrieved. Default is 5")
+      n_samples = gr.Slider(1, 10, value=10, step=1, label="N samples", info="Number of samples to be retrieved. Default is 5")
       threshold = gr.Slider(0.0, 1.0, value=0.75, step=0.01, label="Threshold", info="Threshold of cosine similarity which intent will be considered similar to the input. The higher, the more similar the intent will be. Default is 0.75")
       with gr.Tab("Input from raw text"):
         raw_input_text = gr.Textbox(label="Input Chat", info="Input your chat here, the model will predict the intent")
@@ -188,7 +201,10 @@ with gr.Blocks(title="Intent Classification Demo", theme=theme) as interface:
   accuracy_button.click(fn=check_accuracy, inputs=[n_samples, threshold], outputs=[accuracy_score, accuracy_table])
   raw_ask_button.click(fn=raw_inference, inputs=[raw_input_text, raw_state, n_samples, threshold], outputs=[answer, ask_llm_button_raw])
   ask_button.click(fn=raw_inference, inputs=[input_text, state, n_samples, threshold], outputs=[answer, ask_llm_button])
+  # ask_llm_button.click(fn=classify_intent, inputs=[input_text, history, answer, model_name], outputs=[LLM_prediction, LLM_prompt])
+  # ask_llm_button_raw.click(fn=classify_intent, inputs=[raw_input_text, raw_history, answer, model_name], outputs=[LLM_prediction, LLM_prompt])
   ask_llm_button.click(fn=classify_intent, inputs=[input_text, history, answer, api_key], outputs=[LLM_prediction, LLM_prompt])
   ask_llm_button_raw.click(fn=classify_intent, inputs=[raw_input_text, raw_history, answer, api_key], outputs=[LLM_prediction, LLM_prompt])
 
-interface.launch(debug=True)
+# interface.launch(debug=True)
+interface.launch(share=True, debug=True)

embeddings.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8acf625b84e53b36e293c99f1bccb7e3cb7024357c4489e8d19000c1d0878846
+size 98432

openai_embeddings.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ae474b523b222e95d39016a80eb507b962dc0cebe105744879af1e695d4d456
+size 393344

prompt.py

@@ -1,6 +1,6 @@
 #@title prompt.py
 prompt_template = """
-You are to classify the following message into exactly one of the intents described below:
+The given message needs to be mapped to exactly one of the intents described below. Only answer with the intent name.
 
 {intents}
 unknown: You don't find the matching intent from the above list

utils.py

@@ -38,4 +38,21 @@ def create_embedding(intents:dict, model_en):
   for k,v in intents.items():
     intents_description_en.append(v)
   intents_embedding = model_en.encode(intents_description_en)
-  return intents_embedding
+  return intents_embedding
+
+# def get_embedding(text, model="text-embedding-ada-002"):
+#    text = text.replace("\n", " ")
+#    return client.embeddings.create(input = [text], model=model).data[0].embedding
+ 
+# from openai import OpenAI
+# import numpy as np
+# client = OpenAI()
+
+# def create_embedding_openai(intents:dict):
+#   intents_description_en = []
+#   for k,v in intents.items():
+#     intents_description_en.append(v)
+#   embeddings = np.zeros((len(intents_description_en), 1536))
+#   for i, text in enumerate(intents_description_en):
+#     embeddings[i,:] = get_embedding(text)
+#   return embeddings