Add Template Files

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,14 @@
+---
+title: "Enter Chatbot Title"
+emoji: 📈
+colorFrom: indigo
+colorTo: green
+sdk: streamlit
+sdk_version: 1.32.2
+app_file: app.py
+pinned: false
+license: apache-2.0
+short_description: "Enter Description"
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,93 @@
+from omegaconf import OmegaConf
+from query import VectaraQuery
+import os
+
+import streamlit as st
+from PIL import Image
+from dotenv import load_dotenv
+
+
+load_dotenv(override=False)
+
+def isTrue(x) -> bool:
+    if isinstance(x, bool):
+        return x
+    return x.strip().lower() == 'true'
+
+def launch_bot():
+    def generate_response(question):
+        response = vq.submit_query(question)
+        return response
+    
+    def generate_streaming_response(question):
+        response = vq.submit_query_streaming(question)
+        return response
+
+    if 'cfg' not in st.session_state:
+        corpus_ids = str(os.environ['corpus_ids']).split(',')
+        cfg = OmegaConf.create({
+            'customer_id': str(os.environ['customer_id']),
+            'corpus_ids': corpus_ids,
+            'api_key': str(os.environ['api_key']),
+            'title': os.environ['title'],
+            'description': os.environ['description'],
+            'source_data_desc': os.environ['source_data_desc'],
+            'streaming': isTrue(os.environ.get('streaming', False)),
+            'prompt_name': os.environ.get('prompt_name', None)
+        })
+        st.session_state.cfg = cfg
+        st.session_state.vq = VectaraQuery(cfg.api_key, cfg.customer_id, cfg.corpus_ids, cfg.prompt_name)
+
+    cfg = st.session_state.cfg
+    vq = st.session_state.vq
+    st.set_page_config(page_title=cfg.title, layout="wide")
+
+    # left side content
+    with st.sidebar:
+        image = Image.open('Vectara-logo.png')
+        st.markdown(f"## Welcome to {cfg.title}\n\n"
+                    f"This demo uses Retrieval Augmented Generation to ask questions about {cfg.source_data_desc}\n\n")
+
+        st.markdown("---")
+        st.markdown(
+            "## How this works?\n"
+            "This app was built with [Vectara](https://vectara.com).\n"
+            "Vectara's [Indexing API](https://docs.vectara.com/docs/api-reference/indexing-apis/indexing) was used to ingest the data into a Vectara corpus (or index).\n\n"
+            "This app uses Vectara [Chat API](https://docs.vectara.com/docs/console-ui/vectara-chat-overview) to query the corpus and present the results to you, answering your question.\n\n"
+        )
+        st.markdown("---")
+        st.image(image, width=250)
+
+    st.markdown(f"<center> <h2> Vectara chat demo: {cfg.title} </h2> </center>", unsafe_allow_html=True)
+    st.markdown(f"<center> <h4> {cfg.description} <h4> </center>", unsafe_allow_html=True)
+
+    if "messages" not in st.session_state.keys():
+        st.session_state.messages = [{"role": "assistant", "content": "How may I help you?"}]
+
+    # Display chat messages
+    for message in st.session_state.messages:
+        with st.chat_message(message["role"]):
+            st.write(message["content"])
+
+    # User-provided prompt
+    if prompt := st.chat_input():
+        st.session_state.messages.append({"role": "user", "content": prompt})
+        with st.chat_message("user"):
+            st.write(prompt)
+    
+    # Generate a new response if last message is not from assistant
+    if st.session_state.messages[-1]["role"] != "assistant":
+        with st.chat_message("assistant"):
+            if cfg.streaming:
+                stream = generate_streaming_response(prompt) 
+                response = st.write_stream(stream) 
+            else:
+                with st.spinner("Thinking..."):
+                    response = generate_response(prompt)
+                    st.write(response)
+            message = {"role": "assistant", "content": response}
+            st.session_state.messages.append(message)
+    
+if __name__ == "__main__":
+    launch_bot()
+    

query.py

@@ -0,0 +1,198 @@
+import requests
+import json
+import re
+from urllib.parse import quote
+
+def extract_between_tags(text, start_tag, end_tag):
+    start_index = text.find(start_tag)
+    end_index = text.find(end_tag, start_index)
+    return text[start_index+len(start_tag):end_index-len(end_tag)]
+
+class CitationNormalizer():
+
+    def __init__(self, responses, docs):
+        self.docs = docs
+        self.responses = responses
+        self.refs = []
+
+    def normalize_citations(self, summary):
+        start_tag = "%START_SNIPPET%"
+        end_tag = "%END_SNIPPET%"
+
+        # find all references in the summary
+        pattern = r'\[\d{1,2}\]'
+        matches = [match.span() for match in re.finditer(pattern, summary)]
+
+        # figure out unique list of references
+        for match in matches:
+            start, end = match
+            response_num = int(summary[start+1:end-1])
+            doc_num = self.responses[response_num-1]['documentIndex']
+            metadata = {item['name']: item['value'] for item in self.docs[doc_num]['metadata']}
+            text = extract_between_tags(self.responses[response_num-1]['text'], start_tag, end_tag)
+            if 'url' in metadata.keys():
+                url = f"{metadata['url']}#:~:text={quote(text)}"
+                if url not in self.refs:
+                    self.refs.append(url)
+
+        # replace references with markdown links
+        refs_dict = {url:(inx+1) for inx,url in enumerate(self.refs)}
+        for match in reversed(matches):
+            start, end = match
+            response_num = int(summary[start+1:end-1])
+            doc_num = self.responses[response_num-1]['documentIndex']
+            metadata = {item['name']: item['value'] for item in self.docs[doc_num]['metadata']}
+            text = extract_between_tags(self.responses[response_num-1]['text'], start_tag, end_tag)
+            if 'url' in metadata.keys():
+                url = f"{metadata['url']}#:~:text={quote(text)}"
+                citation_inx = refs_dict[url]
+                summary = summary[:start] + f'[\[{citation_inx}\]]({url})' + summary[end:]
+            else:
+                summary = summary[:start] + summary[end:]
+
+        return summary
+
+class VectaraQuery():
+    def __init__(self, api_key: str, customer_id: str, corpus_ids: list[str], prompt_name: str = None):
+        self.customer_id = customer_id
+        self.corpus_ids = corpus_ids
+        self.api_key = api_key
+        self.prompt_name = prompt_name if prompt_name else "vectara-experimental-summary-ext-2023-12-11-sml"
+        self.conv_id = None
+
+    def get_body(self, query_str: str):
+        corpora_key_list = [{
+                'customer_id': self.customer_id, 'corpus_id': corpus_id, 'lexical_interpolation_config': {'lambda': 0.025}
+            } for corpus_id in self.corpus_ids
+        ]
+
+        return {
+            'query': [
+                { 
+                    'query': query_str,
+                    'start': 0,
+                    'numResults': 50,
+                    'corpusKey': corpora_key_list,
+                    'context_config': {
+                        'sentences_before': 2,
+                        'sentences_after': 2,
+                        'start_tag': "%START_SNIPPET%",
+                        'end_tag': "%END_SNIPPET%",
+                    },
+                    'rerankingConfig':
+                    {
+                        'rerankerId': 272725718,
+                        'mmrConfig': {
+                            'diversityBias': 0.3
+                        }
+                    },
+                    'summary': [
+                        {
+                            'responseLang': 'eng',
+                            'maxSummarizedResults': 5,
+                            'summarizerPromptName': self.prompt_name,
+                            'chat': {
+                                'store': True,
+                                'conversationId': self.conv_id
+                            },
+                        }
+                    ]
+                } 
+            ]
+        }
+
+    def get_headers(self):
+        return {
+            "Content-Type": "application/json",
+            "Accept": "application/json",
+            "customer-id": self.customer_id,
+            "x-api-key": self.api_key,
+            "grpc-timeout": "60S"
+        }
+
+    def submit_query(self, query_str: str):
+
+        endpoint = f"https://api.vectara.io/v1/query"
+        body = self.get_body(query_str)
+
+        response = requests.post(endpoint, data=json.dumps(body), verify=True, headers=self.get_headers())    
+        if response.status_code != 200:
+            print(f"Query failed with code {response.status_code}, reason {response.reason}, text {response.text}")
+            return "Sorry, something went wrong in my brain. Please try again later."
+
+        res = response.json()
+
+        top_k = 10
+        summary = res['responseSet'][0]['summary'][0]['text']
+        responses = res['responseSet'][0]['response'][:top_k]
+        docs = res['responseSet'][0]['document']
+        chat = res['responseSet'][0]['summary'][0].get('chat', None)
+
+        if chat and chat['status'] is not None:
+            st_code = chat['status']
+            print(f"Chat query failed with code {st_code}")
+            if st_code == 'RESOURCE_EXHAUSTED':
+                self.conv_id = None
+                return 'Sorry, Vectara chat turns exceeds plan limit.'
+            return 'Sorry, something went wrong in my brain. Please try again later.'
+        
+        self.conv_id = chat['conversationId'] if chat else None
+        summary = CitationNormalizer(responses, docs).normalize_citations(summary)        
+        return summary
+
+    def submit_query_streaming(self, query_str: str):
+
+        endpoint = f"https://api.vectara.io/v1/stream-query"
+        body = self.get_body(query_str)
+
+        response = requests.post(endpoint, data=json.dumps(body), verify=True, headers=self.get_headers(), stream=True) 
+        if response.status_code != 200:
+            print(f"Query failed with code {response.status_code}, reason {response.reason}, text {response.text}")
+            return "Sorry, something went wrong in my brain. Please try again later."
+
+        chunks = []
+        accumulated_text = ""  # Initialize text accumulation
+        pattern_max_length = 50  # Example heuristic
+        for line in response.iter_lines():
+            if line:  # filter out keep-alive new lines
+                data = json.loads(line.decode('utf-8'))
+                res = data['result']
+                response_set = res['responseSet']                
+                if response_set is None:
+                    # grab next chunk and yield it as output
+                    summary = res.get('summary', None)
+                    if summary is None or len(summary)==0:
+                        continue
+                    else:
+                        chat = summary.get('chat', None)
+                        if chat and chat.get('status', None):
+                            st_code = chat['status']
+                            print(f"Chat query failed with code {st_code}")
+                            if st_code == 'RESOURCE_EXHAUSTED':
+                                self.conv_id = None
+                                return 'Sorry, Vectara chat turns exceeds plan limit.'
+                            return 'Sorry, something went wrong in my brain. Please try again later.'
+                        conv_id = chat.get('conversationId', None) if chat else None
+                        if conv_id:
+                            self.conv_id = conv_id
+                        
+                    chunk = summary['text']
+                    accumulated_text += chunk  # Append current chunk to accumulation
+                    if len(accumulated_text) > pattern_max_length:
+                        accumulated_text = re.sub(r"\[\d+\]", "", accumulated_text)
+                        accumulated_text = re.sub(r"\s+\.", ".", accumulated_text)
+                        out_chunk = accumulated_text[:-pattern_max_length]
+                        chunks.append(out_chunk)
+                        yield out_chunk
+                        accumulated_text = accumulated_text[-pattern_max_length:]
+
+                    if summary['done']:
+                        break
+
+        # yield the last piece
+        if len(accumulated_text) > 0:
+            accumulated_text = re.sub(r" \[\d+\]\.", ".", accumulated_text)
+            chunks.append(accumulated_text)
+            yield accumulated_text        
+        
+        return ''.join(chunks)

requirements.txt

@@ -0,0 +1,4 @@
+requests_to_curl==1.1.0
+toml==0.10.2
+omegaconf==2.3.0
+syrupy==4.0.8