Initial commit

.chroma/chroma-collections.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02c564513891989cdab60ecfe81c42b09048a2d0e33d99b00ff302f9bd1567f2
+size 532

.chroma/chroma-embeddings.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:677722053474bf1aab4c939cf46e3bd7994e14982acf2a02ebaf80418a2e2d8f
+size 1433761

.chroma/index/id_to_uuid_6dc7068f-5504-4150-a2c1-7f17eb4a2ced.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4300adaeb3576278eb1ef7d7ca332603326b53dd4a79465793c69fe3829f7527
+size 3466

.chroma/index/index_6dc7068f-5504-4150-a2c1-7f17eb4a2ced.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:810ec276acabcfb03d602423bf0f8316297a76ebb66600d75d7a6aa0b1b20014
+size 686100

.chroma/index/index_metadata_6dc7068f-5504-4150-a2c1-7f17eb4a2ced.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d262ae2323e12d22ad009e5a8d374fccdffc0c065aec98f4c87d0accddc4264
+size 73

.chroma/index/uuid_to_id_6dc7068f-5504-4150-a2c1-7f17eb4a2ced.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34484296518ecc1d0f9998024eab730a31210b722d21d547400cf8eb6803fb9e
+size 4049

.gitattributes

@@ -0,0 +1,34 @@
+*.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
+*.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

.gitignore

@@ -0,0 +1,195 @@
+# Created by https://www.toptal.com/developers/gitignore/api/python,visualstudiocode
+# Edit at https://www.toptal.com/developers/gitignore?templates=python,visualstudiocode
+
+### Python ###
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/#use-with-ide
+.pdm.toml
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+### Python Patch ###
+# Poetry local configuration file - https://python-poetry.org/docs/configuration/#local-configuration
+poetry.toml
+
+# ruff
+.ruff_cache/
+
+# LSP config files
+pyrightconfig.json
+
+### VisualStudioCode ###
+.vscode/*
+!.vscode/settings.json
+!.vscode/tasks.json
+!.vscode/launch.json
+!.vscode/extensions.json
+!.vscode/*.code-snippets
+
+# Local History for Visual Studio Code
+.history/
+
+# Built Visual Studio Code Extensions
+*.vsix
+
+### VisualStudioCode Patch ###
+# Ignore all local history of files
+.history
+.ionide
+
+# End of https://www.toptal.com/developers/gitignore/api/python,visualstudiocode

README.md

@@ -0,0 +1,12 @@
+---
+title: Doc Search
+emoji: 😻
+colorFrom: pink
+colorTo: green
+sdk: gradio
+sdk_version: 3.32.0
+app_file: app.py
+pinned: false
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,196 @@
+# based on https://github.com/hwchase17/langchain-gradio-template/blob/master/app.py
+import collections
+import os
+from queue import Queue
+from time import sleep
+from typing import Any, Dict, List, Optional, Tuple
+
+import gradio as gr
+from anyio.from_thread import start_blocking_portal
+from langchain import PromptTemplate
+from langchain.callbacks.manager import AsyncCallbackManager
+from langchain.chains import LLMChain
+from langchain.chat_models import ChatOpenAI, PromptLayerChatOpenAI
+from langchain.memory import ConversationBufferMemory
+from langchain.prompts import PromptTemplate
+from langchain.prompts.base import DEFAULT_FORMATTER_MAPPING
+from langchain.prompts.chat import (ChatPromptTemplate,
+                                    HumanMessagePromptTemplate)
+from langchain.schema import HumanMessage
+from langchain.vectorstores import Chroma
+from langchain.docstore.document import Document
+
+from util import SyncStreamingLLMCallbackHandler, CustomOpenAIEmbeddings
+
+
+def I(x):
+    "Identity function; does nothing."
+    return x
+
+class PreprocessingPromptTemplate(PromptTemplate):
+    arg_preprocessing: Dict = {} # this is probably the wrong type
+    def format(self, **kwargs: Any) -> str:
+        """Format the prompt with the inputs.
+
+        Args:
+            kwargs: Any arguments to be passed to the prompt template.
+
+        Returns:
+            A formatted string.
+
+        Example:
+
+        .. code-block:: python
+
+            prompt.format(variable1="foo")
+        """
+        kwargs = self._merge_partial_and_user_variables(**kwargs)
+        kwargs = self._preprocess_args(kwargs)
+        return DEFAULT_FORMATTER_MAPPING[self.template_format](self.template, **kwargs)
+
+    def _preprocess_args(self, args: dict):
+        return {k: self.arg_preprocessing.get(k, I)(v) for k, v in args.items()}
+
+
+def top_results_to_string(x: List[Tuple[Document, float]]):
+    return "\n~~~\n".join(f"Result {i} Title: {doc.metadata['title']}\nResult {i} Content: {doc.page_content}" for i, (doc, score) in enumerate(x, 1))
+
+
+PROMPT = """You are a helpful AI assistant that summarizes search results for users.
+---
+A user has searched for the following query:
+{query}
+---
+The search engine returned the following 5 search results:
+{top_results}
+---
+Based on the search results, answer the user's query, and use the same language as the user's query.
+Say which search result you used.
+Do not use information other than the search results.
+Say 'No answer found.' if there are no relevant results.
+Afterwards, say how confident you are in your answer as a percentage.
+"""
+PROMPT_TEMPLATE = PreprocessingPromptTemplate(template=PROMPT, input_variables=['query', 'top_results'])
+PROMPT_TEMPLATE.arg_preprocessing['top_results'] = top_results_to_string
+
+# TODO give relevance value in prompt
+# TODO ask gpt to say which sources it used
+
+
+
+# TODO azure?
+COLLECTION = Chroma(
+    embedding_function=CustomOpenAIEmbeddings(api_key=os.environ.get("OPENAI_API_KEY", None)),
+    persist_directory="./.chroma",
+    collection_name="CUHK",
+    )
+# COLLECTION = CHROMA_CLIENT.get_collection(name='CUHK')
+
+
+def load_chain(api_type):
+    shared_args = {
+        "temperature": 0,
+        "model_name": "gpt-3.5-turbo",
+        "pl_tags": ["cuhk-demo"],
+        "streaming": True,
+    }
+    if api_type == "OpenAI":
+      chat = PromptLayerChatOpenAI(
+          **shared_args,
+          api_key = os.environ.get("OPENAI_API_KEY", None),
+    )
+    elif api_type == "Azure OpenAI":
+      chat = PromptLayerChatOpenAI(
+        api_type = "azure",
+        api_key = os.environ.get("AZURE_OPENAI_API_KEY", None),
+        api_base = os.environ.get("AZURE_OPENAI_API_BASE", None),
+        api_version = os.environ.get("AZURE_OPENAI_API_VERSION", "2023-03-15-preview"),
+        engine = os.environ.get("AZURE_OPENAI_DEPLOYMENT_NAME", None),
+        **shared_args
+      )
+
+    chain = chain = LLMChain(llm=chat, prompt=PROMPT_TEMPLATE)
+    return chat, chain
+
+def initialize_chain(api_type):
+    "Runs at app start"
+    chat, chain = load_chain(api_type)
+    return chat, chain
+
+def change_chain(api_type, old_chain):
+    chat, chain = load_chain(api_type)
+    return chat, chain
+
+def find_top_results(query):
+    results = COLLECTION.similarity_search_with_score(query, k=4) # TODO filter by device (windows, mac, android, ios)
+
+    output = "\n".join(f"1. [{d.metadata['title']}]({d.metadata['url']}) <small>(dist: {s})</small>" for d, s in results)
+    return results, output
+
+def ask_gpt(chain, query, top_results): # top_results: List[Tuple[Document, float]]
+    q = Queue()
+    job_done = object()
+    def task():
+        chain.run(
+            query=query,
+            top_results=top_results,
+            callbacks=[SyncStreamingLLMCallbackHandler(q)],
+            )
+        q.put(job_done)
+        return
+
+    with start_blocking_portal() as portal:
+        portal.start_task_soon(task)
+
+        content = ""
+        while True:
+            next_token = q.get(True, timeout=15)
+            if next_token is job_done:
+                break
+            content += next_token
+            yield content
+
+
+demo = gr.Blocks(css="""
+#sidebar {
+    max-width: 300px;
+}
+""")
+with demo:
+    with gr.Row():
+        # sidebar
+        with gr.Column(elem_id="sidebar"):
+            api_type = gr.Radio(
+                ["OpenAI", "Azure OpenAI"],
+                value="OpenAI",
+                label="Server",
+                info="You can try changing this if responses are slow."
+            )
+
+        # main
+        with gr.Column():
+            # Company img
+            gr.HTML(r'<div style="display: flex; justify-content: center; align-items: center"><a href="https://thinkcol.com/"><img src="./file=thinkcol-logo.png" alt="ThinkCol" width="357" height="87" /></a></div>')
+
+            chat = gr.State()
+            chain = gr.State()
+
+            query = gr.Textbox(label="Search Query:")
+            top_results_data = gr.State()
+            top_results = gr.Markdown(label="Search Results")
+            response = gr.Textbox(label="AI Response")
+
+
+    load_event = demo.load(initialize_chain, [api_type], [chat, chain])
+
+    query_event = query.submit(find_top_results, [query], [top_results_data, top_results])
+    ask_event = query_event.then(ask_gpt, [chain, query, top_results_data], [response])
+    api_type.change(change_chain,
+                    [api_type, chain],
+                    [chat, chain],
+                    cancels=[load_event, query_event, ask_event])
+
+
+demo.queue()
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1,6 @@
+gradio
+openai
+promptlayer
+langchain
+chromadb
+tiktoken

util.py

@@ -0,0 +1,235 @@
+from typing import Any, Dict, List, Optional, Union
+from types import GeneratorType
+from langchain.callbacks.base import AsyncCallbackHandler, BaseCallbackHandler
+from langchain.schema import AgentAction, AgentFinish, LLMResult
+from langchain.embeddings.openai import embed_with_retry, OpenAIEmbeddings
+from pydantic import Extra, Field, root_validator
+import numpy as np
+
+class StreamingLLMCallbackHandler(AsyncCallbackHandler):
+    """Callback handler for streaming LLM responses to a queue."""
+
+    def __init__(self, q):
+        self.q = q
+
+    def on_llm_new_token(self, token: str, **kwargs: Any) -> None:
+        self.q.put(token)
+
+
+class SyncStreamingLLMCallbackHandler(BaseCallbackHandler):
+    """Callback handler for streaming LLM responses to a queue."""
+
+    def __init__(self, q):
+        self.q = q
+
+    def on_llm_start(
+        self, serialized: Dict[str, Any], prompts: List[str], **kwargs: Any
+    ) -> None:
+        """Do nothing."""
+        pass
+
+    def on_llm_new_token(self, token: str, **kwargs: Any) -> None:
+        self.q.put(token)
+
+    def on_llm_end(self, response: LLMResult, **kwargs: Any) -> None:
+        """Do nothing."""
+        pass
+
+    def on_llm_error(
+        self, error: Union[Exception, KeyboardInterrupt], **kwargs: Any
+    ) -> None:
+        """Do nothing."""
+        pass
+
+    def on_chain_start(
+        self, serialized: Dict[str, Any], inputs: Dict[str, Any], **kwargs: Any
+    ) -> None:
+        """Do nothing."""
+        pass
+
+    def on_chain_end(self, outputs: Dict[str, Any], **kwargs: Any) -> None:
+        """Do nothing."""
+        pass
+
+    def on_chain_error(
+        self, error: Union[Exception, KeyboardInterrupt], **kwargs: Any
+    ) -> None:
+        """Do nothing."""
+        pass
+
+    def on_tool_start(
+        self,
+        serialized: Dict[str, Any],
+        input_str: str,
+        **kwargs: Any,
+    ) -> None:
+        """Do nothing."""
+        pass
+
+    def on_tool_end(
+        self,
+        output: str,
+        color: Optional[str] = None,
+        observation_prefix: Optional[str] = None,
+        llm_prefix: Optional[str] = None,
+        **kwargs: Any,
+    ) -> None:
+        """Do nothing."""
+        pass
+
+    def on_tool_error(
+        self, error: Union[Exception, KeyboardInterrupt], **kwargs: Any
+    ) -> None:
+        """Do nothing."""
+        pass
+
+    def on_agent_action(self, action: AgentAction, **kwargs: Any) -> Any:
+        """Run on agent action."""
+        pass
+
+    def on_agent_finish(
+        self, finish: AgentFinish, color: Optional[str] = None, **kwargs: Any
+    ) -> None:
+        """Run on agent end."""
+        pass
+
+
+def concatenate_generators(*args):
+    final_outputs = ""
+    for g in args:
+        if isinstance(g, GeneratorType):
+            for v in g:
+                yield final_outputs + v
+            result = v
+        else:
+            yield final_outputs + g
+            result = g
+        final_outputs += result
+
+
+class CustomOpenAIEmbeddings(OpenAIEmbeddings):
+    model_kwargs: Dict[str, Any] = Field(default_factory=dict)
+
+    """
+    A version of OpenAIEmbeddings that allows extra args
+    to be passed to OpenAI functions.
+    Based on langchain's ChatOpenAI.
+    """
+    class Config:
+        """Configuration for this pydantic object."""
+
+        extra = Extra.ignore
+
+    @root_validator(pre=True)
+    def build_extra(cls, values: Dict[str, Any]) -> Dict[str, Any]:
+        """Build extra kwargs from additional params that were passed in."""
+        all_required_field_names = {field.alias for field in cls.__fields__.values()}
+
+        extra = values.get("model_kwargs", {})
+        for field_name in list(values):
+            if field_name in extra:
+                raise ValueError(f"Found {field_name} supplied twice.")
+            if field_name not in all_required_field_names:
+                # logger.warning(
+                #     f"""WARNING! {field_name} is not default parameter.
+                #     {field_name} was transferred to model_kwargs.
+                #     Please confirm that {field_name} is what you intended."""
+                # )
+                extra[field_name] = values.pop(field_name)
+
+        disallowed_model_kwargs = all_required_field_names | {"model"}
+        invalid_model_kwargs = disallowed_model_kwargs.intersection(extra.keys())
+        if invalid_model_kwargs:
+            raise ValueError(
+                f"Parameters {invalid_model_kwargs} should be specified explicitly. "
+                f"Instead they were passed in as part of `model_kwargs` parameter."
+            )
+
+        values["model_kwargs"] = extra
+        return values
+
+    # use extra args in calls
+
+    # please refer to
+    # https://github.com/openai/openai-cookbook/blob/main/examples/Embedding_long_inputs.ipynb
+    def _get_len_safe_embeddings(
+        self, texts: List[str], *, engine: str, chunk_size: Optional[int] = None
+    ) -> List[List[float]]:
+        embeddings: List[List[float]] = [[] for _ in range(len(texts))]
+        try:
+            import tiktoken
+
+            tokens = []
+            indices = []
+            encoding = tiktoken.model.encoding_for_model(self.model)
+            for i, text in enumerate(texts):
+                if self.model.endswith("001"):
+                    # See: https://github.com/openai/openai-python/issues/418#issuecomment-1525939500
+                    # replace newlines, which can negatively affect performance.
+                    text = text.replace("\n", " ")
+                token = encoding.encode(
+                    text,
+                    allowed_special=self.allowed_special,
+                    disallowed_special=self.disallowed_special,
+                )
+                for j in range(0, len(token), self.embedding_ctx_length):
+                    tokens += [token[j : j + self.embedding_ctx_length]]
+                    indices += [i]
+
+            batched_embeddings = []
+            _chunk_size = chunk_size or self.chunk_size
+            for i in range(0, len(tokens), _chunk_size):
+                response = embed_with_retry(
+                    self,
+                    input=tokens[i : i + _chunk_size],
+                    engine=self.deployment,
+                    request_timeout=self.request_timeout,
+                    **self.model_kwargs,
+                )
+                batched_embeddings += [r["embedding"] for r in response["data"]]
+
+            results: List[List[List[float]]] = [[] for _ in range(len(texts))]
+            num_tokens_in_batch: List[List[int]] = [[] for _ in range(len(texts))]
+            for i in range(len(indices)):
+                results[indices[i]].append(batched_embeddings[i])
+                num_tokens_in_batch[indices[i]].append(len(tokens[i]))
+
+            for i in range(len(texts)):
+                _result = results[i]
+                if len(_result) == 0:
+                    average = embed_with_retry(
+                        self,
+                        input="",
+                        engine=self.deployment,
+                        request_timeout=self.request_timeout,
+                        **self.model_kwargs,
+                    )["data"][0]["embedding"]
+                else:
+                    average = np.average(
+                        _result, axis=0, weights=num_tokens_in_batch[i]
+                    )
+                embeddings[i] = (average / np.linalg.norm(average)).tolist()
+
+            return embeddings
+
+        except ImportError:
+            raise ValueError(
+                "Could not import tiktoken python package. "
+                "This is needed in order to for OpenAIEmbeddings. "
+                "Please install it with `pip install tiktoken`."
+            )
+
+    def _embedding_func(self, text: str, *, engine: str) -> List[float]:
+        """Call out to OpenAI's embedding endpoint."""
+        # handle large input text
+        if len(text) > self.embedding_ctx_length:
+            return self._get_len_safe_embeddings([text], engine=engine)[0]
+        else:
+            if self.model.endswith("001"):
+                # See: https://github.com/openai/openai-python/issues/418#issuecomment-1525939500
+                # replace newlines, which can negatively affect performance.
+                text = text.replace("\n", " ")
+            return embed_with_retry(
+                self, input=[text], engine=engine, request_timeout=self.request_timeout,
+                **self.model_kwargs,
+            )["data"][0]["embedding"]