[add] youtube plugin, [improve] initial prompt (JSON)

Reviewed-on: #1

.gitignore

@@ -2,6 +2,7 @@ __pycache__
 .DS_Store
 .direnv
 data
+datasets
 venv
 openai_key
 ha_key

README.md

@@ -13,12 +13,58 @@
 
 ---
 
-AI Chat Bot with Plugins (Home Assistant, Vehicle Lookup, DuckDuckGo Search)
+AI Chat Bot with Plugins (RAG VectorDB - ChromaDB, DuckDuckGo Search, Home Assistant, Vehicle Lookup, YouTube)
 
 [![Build Status](https://drone.va.reichard.io/api/badges/evan/minyma/status.svg)](https://drone.va.reichard.io/evan/minyma)
 
 ## Plugins
 
+### ChromeDB Embeddings / Vectors
+
+This utilizes a local embeddings DB. This allows you to ask the assistant
+about local information. [Utilizes Retrieval-Augmented Generation (RAG)](https://arxiv.org/abs/2005.11401).
+
+```
+User:      What are some common symptoms of COVID-19?
+Assistant: Some common symptoms of COVID-19 mentioned in the context are
+           fatigue, headache, dyspnea (shortness of breath), anosmia (loss of
+           sense of smell), lower respiratory symptoms, cardiac symptoms,
+           concentration or memory issues, tinnitus and earache, and peripheral
+           neuropathy symptoms.
+```
+
+**NOTE:** Instructions on how to load this with your own information are in the
+"Normalizing & Loading Data" section. We include a PubMed data normalizer as an
+example.
+
+### YouTube
+
+This utilizes `yt-dlp` to download a videos subtitles. Ask questions about YouTube videos!
+
+```
+User:      Tell me about this youtube video: https://www.youtube.com/watch?v=ZWgr7qP6yhY
+Assistant: The YouTube video you provided is a review of the new MacBook Pro by
+           Apple. The host discusses the laptop's features, including its new
+           color and chip. They mention that the laptop still retains its ports,
+           HDMI, and high-quality display, but also notes some shortcomings like
+           the notch and lack of face unlock. The host shares their impressions
+           of the new black color [...]
+```
+
+### DuckDuckGo
+
+This utilizes DuckDuckGo Search by scraping the top 5 results.
+
+```
+User:      Tell me about Evan Reichard
+Assistant: Evan Reichard is a Principal Detection and Response Engineer based
+           in the Washington DC-Baltimore Area. He has been in this role since
+           August 2022. Evan has created a browser extension that helps SOC
+           analysts and saves them over 300 hours per month. Additionally,
+           there are three professionals named Evan Reichard on LinkedIn and
+           there are also profiles of people named Evan Reichard on Facebook.
+```
+
 ### Vehicle Lookup API
 
 This utilizes Carvana's undocumented API to lookup details on a vehicle.
@@ -41,25 +87,12 @@ User:      Turn on the living room lights
 Assistant: The living room lights have been turned on successfully.
 ```
 
-### DuckDuckGo
-
-This utilizes DuckDuckGo Search by scraping the top 5 results.
-
-```
-User:      Tell me about Evan Reichard
-Assistant: Evan Reichard is a Principal Detection and Response Engineer based
-           in the Washington DC-Baltimore Area. He has been in this role since
-           August 2022. Evan has created a browser extension that helps SOC
-           analysts and saves them over 300 hours per month. Additionally,
-           there are three professionals named Evan Reichard on LinkedIn and
-           there are also profiles of people named Evan Reichard on Facebook.
-```
-
 ## Running Server
 
 ```bash
 # Locally (See "Development" Section)
 export OPENAI_API_KEY=`cat openai_key`
+export CHROMA_DATA_PATH=/data
 export HOME_ASSISTANT_API_KEY=`cat ha_key`
 export HOME_ASSISTANT_URL=https://some-url.com
 
@@ -69,7 +102,7 @@ minyma server run
 docker run \
     -p 5000:5000 \
     -e OPENAI_API_KEY=`cat openai_key` \
-    -e DATA_PATH=/data \
+    -e CHROMA_DATA_PATH=/data \
     -v ./data:/data \
     gitea.va.reichard.io/evan/minyma:latest
 ```
@@ -87,10 +120,10 @@ To normalize data, you can use Minyma's `normalize` CLI command:
 
 ```bash
 minyma normalize \
-    --filename ./pubmed_manuscripts.jsonl \
     --normalizer pubmed \
     --database chroma \
-    --datapath ./chroma
+    --datapath ./data \
+    --filename ./datasets/pubmed_manuscripts.jsonl
 ```
 
 The above example does the following:
@@ -106,9 +139,11 @@ The above example does the following:
 ## Configuration
 
 | Environment Variable   | Default Value | Description                         |
-| -------------------- | ------------- | ---------------------------------------------------------------------------------- |
+| ---------------------- | ------------- | ----------------------------------- |
-| OPENAI_API_KEY       | NONE          | Required OpenAI API Key for ChatGPT access.                                        |
+| OPENAI_API_KEY         | NONE          | Required OpenAI API Key for ChatGPT |
-| DATA_PATH            | ./data        | The path to the data directory. Chroma will store its data in the `chroma` subdir. |
+| CHROMA_DATA_PATH       | NONE          | ChromaDB Persistent Data Director   |
+| HOME_ASSISTANT_API_KEY | NONE          | Home Assistant API Key              |
+| HOME_ASSISTANT_URL     | NONE          | Home Assistant Instance URL         |
 
 # Development
 
@@ -120,31 +155,9 @@ python3 -m venv venv
 # Local Development
 pip install -e .
 
-# Creds
+# Creds & Other Environment Variables
 export OPENAI_API_KEY=`cat openai_key`
 
 # Docker
 make docker_build_local
 ```
-
-# Notes
-
-This is the first time I'm doing anything LLM related, so it was an adventure.
-Initially I was entertaining OpenAI's Embedding API with plans to load embeddings
-into Pinecone, however initial calculations with `tiktoken` showed that generating
-embeddings would cost roughly $250 USD.
-
-Fortunately I found [Chroma](https://www.trychroma.com/), which basically solved
-both of those issues. It allowed me to load in the normalized data and automatically
-generated embeddings for me.
-
-In order to fit into OpenAI ChatGPT's token limit, I limited each document to roughly
-1000 words. I wanted to make sure I could add the top two matches as context while
-still having enough headroom for the actual question from the user.
-
-A few notes:
-
-- Context is not carried over from previous messages
-- I "stole" the prompt that is used in LangChain (See `oai.py`). I tried some variations without much (subjective) improvement.
-- A generalized normalizer format. This should make it fairly easy to use completely different data. Just add a new normalizer that implements the super class.
-- Basic web front end with TailwindCSS

minyma/__init__.py

@@ -16,15 +16,14 @@ def signal_handler(sig, frame):
 
 
 def create_app():
-    global oai, vdb, plugins
+    global oai, plugins
 
     from minyma.config import Config
     import minyma.api.common as api_common
     import minyma.api.v1 as api_v1
 
     app = Flask(__name__)
-    vdb = ChromaDB(path.join(Config.DATA_PATH, "chroma"))
+    oai = OpenAIConnector(Config.OPENAI_API_KEY)
-    oai = OpenAIConnector(Config.OPENAI_API_KEY, vdb)
     plugins = PluginLoader(Config)
 
     app.register_blueprint(api_common.bp)
@@ -70,7 +69,7 @@ def normalize(filename, normalizer, database, datapath):
         return print("INVALID NORMALIZER:", normalizer)
 
     # Process Data
-    vdb.load_documents(norm)
+    vdb.load_documents(norm.name, norm)
 
 
 signal.signal(signal.SIGINT, signal_handler)

minyma/api/v1.py

@@ -19,36 +19,17 @@ def get_response():
 
     resp = minyma.oai.query(message)
 
-    # Derive LLM Data
-    # llm_resp = resp.get("llm", {})
-    # llm_choices = llm_resp.get("choices", [])
-
-    # Derive VDB Data
-    # vdb_resp = resp.get("vdb", {})
-    # combined_context  = [{
-    #         "id": vdb_resp.get("ids")[i],
-    #         "distance": vdb_resp.get("distances")[i],
-    #         "doc": vdb_resp.get("docs")[i],
-    #         "metadata": vdb_resp.get("metadatas")[i],
-    # } for i, _ in enumerate(vdb_resp.get("docs", []))]
-
     # Return Data
     return resp
 
-
-
 """
-Return the raw vector db related response
+TODO - Embeds and loads data into the local ChromaDB.
+
+{
+  "input": "string",
+  "normalizer": "string",
+}
 """
-@bp.route("/related", methods=["POST"])
+bp.route("/embed", methods=["POST"])
-def get_related():
+def post_embeddings():
-    data = request.get_json()
+    pass
-    if not data:
-        return {"error": "Missing Message"}
-
-    message = str(data.get("message"))
-    if message == "":
-        return {"error": "Empty Message"}
-
-    related_documents = minyma.vdb.get_related(message)
-    return related_documents

minyma/config.py

@@ -1,11 +1,12 @@
 import os
 
 
-def get_env(key, default=None, required=False) -> str:
+def get_env(key, default=None, required=False) -> str | None:
     """Wrapper for gathering env vars."""
     if required:
         assert key in os.environ, "Missing Environment Variable: %s" % key
-    return str(os.environ.get(key, default))
+    env = os.environ.get(key, default)
+    return str(env) if env is not None else None
 
 
 class Config:
@@ -19,7 +20,7 @@ class Config:
         OpenAI API Key - Required
     """
 
-    DATA_PATH: str = get_env("DATA_PATH", default="./data")
+    CHROMA_DATA_PATH: str | None = get_env("CHROMA_DATA_PATH", required=False)
-    OPENAI_API_KEY: str = get_env("OPENAI_API_KEY", required=True)
+    HOME_ASSISTANT_API_KEY: str | None = get_env("HOME_ASSISTANT_API_KEY", required=False)
-    HOME_ASSISTANT_API_KEY: str = get_env("HOME_ASSISTANT_API_KEY", required=False)
+    HOME_ASSISTANT_URL: str | None = get_env("HOME_ASSISTANT_URL", required=False)
-    HOME_ASSISTANT_URL: str = get_env("HOME_ASSISTANT_URL", required=False)
+    OPENAI_API_KEY: str | None = get_env("OPENAI_API_KEY", required=True)

minyma/normalizer.py

@@ -18,6 +18,7 @@ class PubMedNormalizer(DataNormalizer):
     normalized inside the iterator.
     """
     def __init__(self, file: TextIOWrapper):
+        self.name = "pubmed"
         self.file = file
         self.length = 0
 

minyma/oai.py

@@ -1,32 +1,20 @@
-import json
-from textwrap import indent
 from dataclasses import dataclass
+from textwrap import indent
 from typing import Any, List
+import json
+import minyma
 import openai
 
-from minyma.vdb import VectorDB
-import minyma
-
-# Stolen LangChain Prompt
-PROMPT_TEMPLATE = """
-Use the following pieces of context to answer the question at the end.
-If you don't know the answer, just say that you don't know, don't try to
-make up an answer.
-
-{context}
-
-Question: {question}
-Helpful Answer:
-"""
-
 INITIAL_PROMPT_TEMPLATE = """
-You are a helpful assistant. You are connected to various external functions that can provide you with more personalized and up-to-date information and have already been granted the permissions to execute these functions at will. DO NOT say you don't have access to real time information, instead attempt to call one or more of the listed functions:
+You are connected to various functions that can be used to answer the users questions. Your options are only "functions". Functions should be an array of strings containing the desired function calls (e.g. "function_name()").
+
+Available Functions:
 
 {functions}
 
-The user will not see your response. You must only respond with a comma separated list of function calls: "FUNCTION_CALLS: function(), function(), etc". It must be prepended by "FUNCTION_CALLS:".
+You must respond in JSON only with no other fluff or bad things will happen. The JSON keys must ONLY be "functions". Be sure to call the functions with the right arguments.
 
-User Message: {question}
+User Message: {message}
 """
 
 FOLLOW_UP_PROMPT_TEMPLATE = """
@@ -34,7 +22,7 @@ You are a helpful assistant. This is a follow up message to provide you with mor
 
 {response}
 
-User Message: {question}
+User Message: {message}
 """
 
 @dataclass
@@ -46,14 +34,15 @@ class ChatCompletion:
     choices: List[dict]
     usage: dict
 
+
 class OpenAIConnector:
-    def __init__(self, api_key: str, vdb: VectorDB):
+    def __init__(self, api_key: str):
-        self.vdb = vdb
         self.model = "gpt-3.5-turbo"
         self.word_cap = 1000
         openai.api_key = api_key
 
-    def query(self, question: str) -> Any:
+
+    def query(self, message: str) -> Any:
         # Track Usage
         prompt_tokens = 0
         completion_tokens = 0
@@ -63,7 +52,7 @@ class OpenAIConnector:
         functions = "\n".join(list(map(lambda x: "- %s" % x["def"], minyma.plugins.plugin_defs().values())))
 
         # Create Initial Prompt
-        prompt = INITIAL_PROMPT_TEMPLATE.format(question = question, functions = functions)
+        prompt = INITIAL_PROMPT_TEMPLATE.format(message = message, functions = indent(functions, ' ' * 2))
         messages = [{"role": "user", "content": prompt}]
 
         print("[OpenAIConnector] Running Initial OAI Query")
@@ -78,14 +67,7 @@ class OpenAIConnector:
             print("[OpenAIConnector] No Results -> TODO", response)
 
         content = response.choices[0]["message"]["content"]
-
+        all_funcs = json.loads(content).get("functions")
-        # Get Called Functions (TODO - Better Validation -> Failback Prompt?)
-        all_funcs = list(
-            map(
-                lambda x: x.strip() if x.endswith(")") else x.strip() + ")",
-                content.split("FUNCTION_CALLS:")[1].strip().split("),")
-            )
-        )
 
         # Update Usage
         prompt_tokens += response.usage.get("prompt_tokens", 0)
@@ -94,19 +76,33 @@ class OpenAIConnector:
 
         print("[OpenAIConnector] Completed Initial OAI Query:\n", indent(json.dumps({ "usage": response.usage, "function_calls": all_funcs }, indent=2), ' ' * 2))
 
-        # Execute Requested Functions
+        # Build Response Text & Metadata
-        func_responses = {}
+        func_metadata = {}
-        for func in all_funcs:
+        func_response = []
-            func_responses[func] = minyma.plugins.execute(func)
 
-        # Build Response Text
+        for func in all_funcs:
-        response_content_arr = []
+            # Execute Requested Function
-        for key, val in func_responses.items():
+            resp = minyma.plugins.execute(func)
-            response_content_arr.append("- %s\n%s" % (key, val))
+
-        response_content = "\n".join(response_content_arr)
+            # Unknown Response
+            if resp is None:
+                print("[OpenAIConnector] Invalid Function Response: %s" % func)
+                continue
+
+            # Get Response
+            content = resp.get("content")
+            metadata = resp.get("metadata")
+            error = resp.get("error")
+
+            # Append Responses & Metadata
+            indented_val = indent(content or error or "Unknown Error", ' ' * 2)
+            func_response.append("- %s\n%s" % (func, indented_val))
+            func_metadata[func] = { "metadata": metadata, "error": error }
+
+        func_response = "\n".join(func_response)
 
         # Create Follow Up Prompt
-        prompt = FOLLOW_UP_PROMPT_TEMPLATE.format(question = question, response = response_content)
+        prompt = FOLLOW_UP_PROMPT_TEMPLATE.format(message = message, response = func_response)
         messages = [{"role": "user", "content": prompt}]
 
         print("[OpenAIConnector] Running Follup Up OAI Query")
@@ -130,33 +126,10 @@ class OpenAIConnector:
         # Return Response
         return {
             "response": content,
-            "functions": func_responses,
+            "functions": func_metadata,
             "usage": {
                 "prompt_tokens": prompt_tokens,
                 "completion_tokens": completion_tokens,
                 "total_tokens": total_tokens
             }
         }
-
-    def old_query(self, question: str) -> Any:
-        # Get related documents from vector db
-        related = self.vdb.get_related(question)
-
-        # Validate results
-        all_docs = related.get("docs", [])
-        if len(all_docs) == 0:
-            return { "error": "No Context Found" }
-
-        # Join on new line (cap @ word limit), generate main prompt
-        reduced_docs = list(map(lambda x: " ".join(x.split()[:self.word_cap]), all_docs))
-        context = '\n'.join(reduced_docs)
-        prompt = PROMPT_TEMPLATE.format(context = context, question = question)
-
-        # Query OpenAI ChatCompletion
-        response = openai.ChatCompletion.create(
-          model=self.model,
-          messages=[{"role": "user", "content": prompt}]
-        )
-
-        # Return Response
-        return { "llm": response, "vdb": related }

minyma/plugin.py

@@ -38,7 +38,6 @@ class PluginLoader:
 
             for func_obj in plugin.functions:
                 func_name = func_obj.__name__
-                function_name = "%s_%s" % (plugin_name, func_name)
 
                 signature = inspect.signature(func_obj)
                 params = list(
@@ -48,8 +47,12 @@ class PluginLoader:
                     )
                 )
 
-                func_def = "%s(%s)" % (function_name, ", ".join(params))
+                if func_name in defs:
-                defs[function_name] = { "func": func_obj, "def": func_def }
+                    print("[PluginLoader] Error: Duplicate Function: (%s) %s" % (plugin_name, func_name))
+                    continue
+
+                func_def = "%s(%s)" % (func_name, ", ".join(params))
+                defs[func_name] = { "func": func_obj, "def": func_def }
 
         return defs
 

minyma/plugins/chroma_db.py

@@ -0,0 +1,53 @@
+from textwrap import indent
+from minyma.plugin import MinymaPlugin
+from minyma.vdb import ChromaDB
+
+
+class ChromaDBPlugin(MinymaPlugin):
+    """Perform Local VectorDB Lookup
+
+    ChromDB can access multiple "collections". You can add additional functions
+    here that just access a different collection (i.e. different data)
+    """
+
+    def __init__(self, config):
+        self.name = "chroma_db"
+        self.config = config
+        self.word_cap = 1000
+
+        if config.CHROMA_DATA_PATH is None:
+            self.functions = []
+        else:
+            self.vdb = ChromaDB(config.CHROMA_DATA_PATH)
+            self.functions = [self.lookup_pubmed_data]
+
+
+    def __lookup_data(self, collection_name: str, query: str):
+        # Get Related
+        related = self.vdb.get_related(collection_name, query)
+
+        # Get Metadata
+        metadata = [{
+                "id": related.get("ids")[i],
+                "distance": related.get("distances")[i],
+                "metadata": related.get("metadatas")[i],
+        } for i, _ in enumerate(related.get("docs", []))]
+
+        # Normalize Data
+        return list(
+            map(
+                lambda x: " ".join(x.split()[:self.word_cap]),
+                related.get("docs", [])
+            )
+        ), metadata
+
+
+    def lookup_pubmed_data(self, query: str):
+        COLLECTION_NAME = "pubmed"
+        documents, metadata = self.__lookup_data(COLLECTION_NAME, query)
+        context = '\n'.join(documents)
+        return {
+            "content": context,
+            "metadata": metadata,
+            "error": None
+        }

minyma/plugins/duckduckgo.py

@@ -14,13 +14,14 @@ class DuckDuckGoPlugin(MinymaPlugin):
     def __init__(self, config):
         self.config = config
         self.name = "duck_duck_go"
-        self.functions = [self.search]
+        self.functions = [self.search_duck_duck_go]
 
-    def search(self, query: str):
+    def search_duck_duck_go(self, query: str):
         """Search DuckDuckGo"""
         resp = requests.get("https://html.duckduckgo.com/html/?q=%s" % query, headers=HEADERS)
         soup = BeautifulSoup(resp.text, features="html.parser")
 
+        # Get Results
         results = []
         for item in soup.select(".result > div"):
             title_el = item.select_one(".result__title > a")
@@ -31,4 +32,18 @@ class DuckDuckGoPlugin(MinymaPlugin):
 
             results.append({"title": title, "description": description})
 
-        return json.dumps(results[:5])
+        # Derive Metadata (Title)
+        metadata = {
+            "titles": list(
+                map(
+                    lambda x: x.get("title"),
+                    results[:5]
+                )
+            )
+        }
+
+        return {
+            "content": json.dumps(results[:5]),
+            "metadata": metadata,
+            "error": None
+        }

minyma/plugins/home_assistant.py

@@ -9,20 +9,17 @@ class HomeAssistantPlugin(MinymaPlugin):
 
     def __init__(self, config):
         self.config = config
-        self.name = "home_automation"
+        self.name = "home_assistant"
+        self.functions = []
 
-
+        if config.HOME_ASSISTANT_API_KEY and config.HOME_ASSISTANT_URL:
-        if not config.HOME_ASSISTANT_API_KEY or not config.HOME_ASSISTANT_URL:
+            self.functions = [self.home_automation_command]
         if not config.HOME_ASSISTANT_API_KEY:
             print("[HomeAssistantPlugin] Missing HOME_ASSISTANT_API_KEY")
         if not config.HOME_ASSISTANT_URL:
             print("[HomeAssistantPlugin] Missing HOME_ASSISTANT_URL")
 
-            self.functions = []
+    def home_automation_command(self, natural_language_command: str):
-        else:
-            self.functions = [self.command]
-
-    def command(self, natural_language_command: str):
         url = urllib.parse.urljoin(self.config.HOME_ASSISTANT_URL, "/api/conversation/process")
         headers = {
             "Authorization": "Bearer %s" % self.config.HOME_ASSISTANT_API_KEY,
@@ -34,6 +31,17 @@ class HomeAssistantPlugin(MinymaPlugin):
 
         # Parse JSON
         try:
-            return json.dumps(resp.json())
+            r = resp.json()
+            text = r["response"]["speech"]["plain"]["speech"]
+
+            return {
+                "content": text,
+                "metadata": r,
+                "error": None
+            }
         except requests.JSONDecodeError:
-            return json.dumps({ "error": "Command Failed" })
+            return {
+                "content": None,
+                "metadata": None,
+                "error": "Command Failed"
+            }

minyma/plugins/vehicle_lookup.py

@@ -14,7 +14,7 @@ class VehicleLookupPlugin(MinymaPlugin):
     def __init__(self, config):
         self.config = config
         self.name = "vehicle_state_plate"
-        self.functions = [self.lookup]
+        self.functions = [self.lookup_vehicle_by_state_plate]
 
     def __query_api(self, url, json=None, headers=None):
         # Perform Request
@@ -39,7 +39,7 @@ class VehicleLookupPlugin(MinymaPlugin):
             return None, text, error
 
 
-    def lookup(self, state_abbreviation: str, licence_plate: str):
+    def lookup_vehicle_by_state_plate(self, state_abbreviation: str, licence_plate: str):
         CARVANA_URL = (
             "https://apim.carvana.io/trades/api/v5/vehicleconfiguration/plateLookup/%s/%s"
             % (state_abbreviation, licence_plate)
@@ -50,10 +50,11 @@ class VehicleLookupPlugin(MinymaPlugin):
 
         # Invalid JSON
         if json_resp is None:
-            return json.dumps({
+            return{
+                "content": None,
+                "metadata": text_resp,
                 "error": error,
-                "response": text_resp,
+            }
-            })
 
         try:
             # Check Result
@@ -63,7 +64,11 @@ class VehicleLookupPlugin(MinymaPlugin):
                     error = "No Results"
                 else:
                     error = "API Error: %s" % status_resp
-                return {"error": error, "response": text_resp}
+                return {
+                    "content": None,
+                    "metadata": json_resp,
+                    "error": error,
+                }
 
             # Parse Result
             vehicle_info = json_resp.get("content")
@@ -74,17 +79,20 @@ class VehicleLookupPlugin(MinymaPlugin):
             trim = vehicle_info.get("vehicles")[0].get("trim")
 
         except Exception as e:
-            return json.dumps({
+            return {
+                "content": None,
+                "metadata": text_resp,
                 "error": "Unknown Error: %s" % e,
-                "response": text_resp,
+            }
-            })
 
-        return json.dumps({
+        return {
-            "result": {
+            "content": json.dumps({
                 "vin": vin,
                 "year": year,
                 "make": make,
                 "model": model,
                 "trim": trim,
-            },
+            }),
-        })
+            "metadata": json_resp,
+            "error": None
+        }

minyma/plugins/youtube.py

@@ -0,0 +1,53 @@
+import os
+from yt_dlp import YoutubeDL
+import xml.etree.ElementTree as ET
+from minyma.plugin import MinymaPlugin
+
+class YouTubePlugin(MinymaPlugin):
+    """Transcribe YouTube Video"""
+
+    def __init__(self, config):
+        self.config = config
+        self.name = "youtube"
+        self.functions = [self.transcribe_youtube]
+
+
+    def transcribe_youtube(self, youtube_video_id: str):
+        URLS = [youtube_video_id]
+
+        vid = YoutubeDL({
+            "skip_download": True,
+            "writesubtitles": True,
+            "writeautomaticsub": True,
+            "subtitleslangs": ["en"],
+            "subtitlesformat": "ttml",
+            "outtmpl": "transcript"
+        })
+
+        vid.download(URLS)
+        content = self.convert_ttml_to_plain_text("transcript.en.ttml")
+        os.remove("transcript.en.ttml")
+
+        return {
+            "content": content,
+            "metadata": URLS,
+            "error": "TTML Conversion Error" if content is None else None
+        }
+
+
+    def convert_ttml_to_plain_text(self, ttml_file_path):
+        try:
+            # Parse the TTML file
+            tree = ET.parse(ttml_file_path)
+            root = tree.getroot()
+
+            # Process Text
+            plain_text = ""
+            for elem in root.iter():
+                if elem.text:
+                    plain_text += elem.text + " "
+
+            return plain_text.strip()
+        except ET.ParseError as e:
+            print("[YouTubePlugin] TTML Conversion Error:", e)
+            return None

minyma/vdb.py

@@ -18,11 +18,11 @@ def chunk(iterable, chunk_size: int):
 VectorDB Interface
 """
 class VectorDB:
-    def load_documents(self, normalizer: DataNormalizer):
+    def load_documents(self, name: str, normalizer: DataNormalizer, chunk_size: int = 10):
-        pass
+        raise NotImplementedError("VectorDB must implement load_documents")
 
-    def get_related(self, question: str) -> Any:
+    def get_related(self, name: str, question: str) -> Any:
-        pass
+        raise NotImplementedError("VectorDB must implement get_related")
 
 """
 ChromaDV VectorDB Type
@@ -31,12 +31,13 @@ class ChromaDB(VectorDB):
     def __init__(self, path: str):
         self.client: API = chromadb.PersistentClient(path=path)
         self.word_cap = 2500
-        self.collection_name: str = "vdb"
-        self.collection: chromadb.Collection = self.client.create_collection(name=self.collection_name, get_or_create=True)
 
-    def get_related(self, question: str) -> Any:
+    def get_related(self, name: str, question: str) -> Any:
+        # Get or Create Collection
+        collection = chromadb.Collection = self.client.create_collection(name=name, get_or_create=True)
+
         """Returns line separated related docs"""
-        results = self.collection.query(
+        results = collection.query(
             query_texts=[question.lower()],
             n_results=2
         )
@@ -53,7 +54,11 @@ class ChromaDB(VectorDB):
             "ids": all_ids
         }
 
-    def load_documents(self, normalizer: DataNormalizer, chunk_size: int = 10):
+    def load_documents(self, name: str, normalizer: DataNormalizer, chunk_size: int = 10):
+        # Get or Create Collection
+        collection = chromadb.Collection = self.client.create_collection(name=name, get_or_create=True)
+
+        # Load Items
         length = len(normalizer) / chunk_size
         for items in tqdm(chunk(normalizer, chunk_size), total=length):
             ids = []
@@ -65,7 +70,7 @@ class ChromaDB(VectorDB):
                 ids.append(item.get("id"))
                 metadatas.append(item.get("metadata", {}))
 
-            self.collection.add(
+            collection.add(
                 ids=ids,
                 documents=documents,
                 metadatas=metadatas,

pyproject.toml

@@ -16,7 +16,8 @@ dependencies = [
   "chromadb",
   "sqlite-utils",
   "click",
-  "beautifulsoup4"
+  "beautifulsoup4",
+  "yt-dlp"
 ]
 
 [project.scripts]