Upload folder using huggingface_hub
Browse files- app.py +66 -25
- main.py +2 -2
- modules/collector.py +47 -51
- modules/db.py +3 -3
app.py
CHANGED
|
@@ -3,7 +3,7 @@ import re
|
|
| 3 |
import gradio as gr
|
| 4 |
from gradio_modal import Modal
|
| 5 |
import chromadb
|
| 6 |
-
from modules.collector import
|
| 7 |
from modules.db import (
|
| 8 |
delete_channel_from_collection,
|
| 9 |
get_collection,
|
|
@@ -62,11 +62,23 @@ def enable_if_not_none(question):
|
|
| 62 |
# Fetch & index channels
|
| 63 |
# -------------------------------
|
| 64 |
def refresh_channel(api_key, channel_url: str):
|
| 65 |
-
|
| 66 |
-
for
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 70 |
|
| 71 |
|
| 72 |
def index_channels(channel_urls: str):
|
|
@@ -111,31 +123,59 @@ def list_channels_radio():
|
|
| 111 |
channel_id = key
|
| 112 |
if channel_id:
|
| 113 |
choices.append((channel_display_name, channel_id))
|
| 114 |
-
print("choices= ", choices)
|
| 115 |
return choices
|
| 116 |
|
| 117 |
|
| 118 |
# -------------------------------
|
| 119 |
# Fetch channel videos as HTML table
|
| 120 |
# -------------------------------
|
| 121 |
-
def fetch_channel_html(
|
| 122 |
-
|
| 123 |
-
|
| 124 |
-
|
| 125 |
-
|
| 126 |
-
|
| 127 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 128 |
for idx, v in enumerate(videos):
|
| 129 |
-
html += "
|
| 130 |
-
|
| 131 |
-
|
| 132 |
-
|
| 133 |
-
|
| 134 |
-
|
| 135 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 136 |
return html
|
| 137 |
|
| 138 |
-
get_collection # -------------------------------
|
| 139 |
|
| 140 |
|
| 141 |
# Delete a channel
|
|
@@ -203,7 +243,7 @@ with gr.Blocks() as demo:
|
|
| 203 |
with gr.Row():
|
| 204 |
gr.Column()
|
| 205 |
save_add_channels_btn = gr.Button(
|
| 206 |
-
"Add
|
| 207 |
)
|
| 208 |
gr.Column()
|
| 209 |
index_status = gr.Markdown(label="Index Status", container=False)
|
|
@@ -358,7 +398,7 @@ with gr.Blocks() as demo:
|
|
| 358 |
|
| 359 |
# Show videos modal when button clicked
|
| 360 |
def show_selected_channel_videos(selected_channel_id):
|
| 361 |
-
print("selected_channel_id = ", selected_channel_id)
|
| 362 |
return fetch_channel_html(selected_channel_id)
|
| 363 |
|
| 364 |
channel_radio.change(
|
|
@@ -395,7 +435,8 @@ with gr.Blocks() as demo:
|
|
| 395 |
def init():
|
| 396 |
channels = "https://www.youtube.com/@onedayonepasuram6126,https://www.youtube.com/@srisookthi,https://www.youtube.com/@learn-aksharam"
|
| 397 |
for resp in index_channels(channels):
|
| 398 |
-
print(resp)
|
|
|
|
| 399 |
|
| 400 |
|
| 401 |
if __name__ == "__main__":
|
|
|
|
| 3 |
import gradio as gr
|
| 4 |
from gradio_modal import Modal
|
| 5 |
import chromadb
|
| 6 |
+
from modules.collector import fetch_all_channel_videos
|
| 7 |
from modules.db import (
|
| 8 |
delete_channel_from_collection,
|
| 9 |
get_collection,
|
|
|
|
| 62 |
# Fetch & index channels
|
| 63 |
# -------------------------------
|
| 64 |
def refresh_channel(api_key, channel_url: str):
|
| 65 |
+
all_videos = []
|
| 66 |
+
for status, videos_batch in fetch_all_channel_videos(api_key, channel_url):
|
| 67 |
+
# enrich each video in the batch with channel_url
|
| 68 |
+
for v in videos_batch:
|
| 69 |
+
v["channel_url"] = channel_url
|
| 70 |
+
all_videos.extend(videos_batch)
|
| 71 |
+
|
| 72 |
+
# (optional) log progress
|
| 73 |
+
print(status, "- total so far:", len(all_videos))
|
| 74 |
+
|
| 75 |
+
# index all at once
|
| 76 |
+
if all_videos:
|
| 77 |
+
print(f"Adding {len(all_videos)} videos to database")
|
| 78 |
+
index_videos(all_videos, get_collection(), channel_url=channel_url)
|
| 79 |
+
print(f"Loaded {len(all_videos)} videos")
|
| 80 |
+
|
| 81 |
+
return len(all_videos)
|
| 82 |
|
| 83 |
|
| 84 |
def index_channels(channel_urls: str):
|
|
|
|
| 123 |
channel_id = key
|
| 124 |
if channel_id:
|
| 125 |
choices.append((channel_display_name, channel_id))
|
| 126 |
+
# print("choices= ", choices)
|
| 127 |
return choices
|
| 128 |
|
| 129 |
|
| 130 |
# -------------------------------
|
| 131 |
# Fetch channel videos as HTML table
|
| 132 |
# -------------------------------
|
| 133 |
+
def fetch_channel_html(channel_id: str):
|
| 134 |
+
# query your collection/db instead of YouTube API
|
| 135 |
+
collection = get_collection()
|
| 136 |
+
results = collection.get(
|
| 137 |
+
where={"channel_id": channel_id},
|
| 138 |
+
include=["documents", "metadatas"]
|
| 139 |
+
)
|
| 140 |
+
|
| 141 |
+
if not results or not results.get("metadatas"):
|
| 142 |
+
return """
|
| 143 |
+
<div style="display:flex;justify-content:center;align-items:center;
|
| 144 |
+
height:200px;flex-direction:column;color:#666;">
|
| 145 |
+
⚠️ No videos found for this channel.
|
| 146 |
+
</div>
|
| 147 |
+
"""
|
| 148 |
+
|
| 149 |
+
videos = results["metadatas"]
|
| 150 |
+
|
| 151 |
+
# build table
|
| 152 |
+
html = """
|
| 153 |
+
<table border="1" style="border-collapse:collapse;width:100%;font-family:sans-serif;">
|
| 154 |
+
<thead style="background:#f0f0f0;">
|
| 155 |
+
<tr>
|
| 156 |
+
<th>#</th>
|
| 157 |
+
<th>Title</th>
|
| 158 |
+
<th>Video URL</th>
|
| 159 |
+
<th>Description</th>
|
| 160 |
+
</tr>
|
| 161 |
+
</thead>
|
| 162 |
+
<tbody>
|
| 163 |
+
"""
|
| 164 |
+
|
| 165 |
for idx, v in enumerate(videos):
|
| 166 |
+
html += f"""
|
| 167 |
+
<tr>
|
| 168 |
+
<td>{idx+1}</td>
|
| 169 |
+
<td>{v.get('video_title','')}</td>
|
| 170 |
+
<td><a href="https://youtube.com/watch?v={v.get('video_id')}"
|
| 171 |
+
target="_blank">Watch Video</a></td>
|
| 172 |
+
<td>{v.get('description','')}</td>
|
| 173 |
+
</tr>
|
| 174 |
+
"""
|
| 175 |
+
|
| 176 |
+
html += "</tbody></table>"
|
| 177 |
return html
|
| 178 |
|
|
|
|
| 179 |
|
| 180 |
|
| 181 |
# Delete a channel
|
|
|
|
| 243 |
with gr.Row():
|
| 244 |
gr.Column()
|
| 245 |
save_add_channels_btn = gr.Button(
|
| 246 |
+
"Add Channel(s)", scale=0, variant="primary"
|
| 247 |
)
|
| 248 |
gr.Column()
|
| 249 |
index_status = gr.Markdown(label="Index Status", container=False)
|
|
|
|
| 398 |
|
| 399 |
# Show videos modal when button clicked
|
| 400 |
def show_selected_channel_videos(selected_channel_id):
|
| 401 |
+
# print("selected_channel_id = ", selected_channel_id)
|
| 402 |
return fetch_channel_html(selected_channel_id)
|
| 403 |
|
| 404 |
channel_radio.change(
|
|
|
|
| 435 |
def init():
|
| 436 |
channels = "https://www.youtube.com/@onedayonepasuram6126,https://www.youtube.com/@srisookthi,https://www.youtube.com/@learn-aksharam"
|
| 437 |
for resp in index_channels(channels):
|
| 438 |
+
# print(resp)
|
| 439 |
+
pass
|
| 440 |
|
| 441 |
|
| 442 |
if __name__ == "__main__":
|
main.py
CHANGED
|
@@ -10,7 +10,7 @@ import chromadb
|
|
| 10 |
from dotenv import load_dotenv
|
| 11 |
|
| 12 |
from modules.answerer import answer_query
|
| 13 |
-
from modules.collector import
|
| 14 |
from modules.db import get_collection
|
| 15 |
from modules.indexer import index_videos
|
| 16 |
|
|
@@ -26,7 +26,7 @@ def main():
|
|
| 26 |
|
| 27 |
# Collect + Index
|
| 28 |
for ch in CHANNELS:
|
| 29 |
-
videos =
|
| 30 |
index_videos(videos, collection)
|
| 31 |
|
| 32 |
# Ask a question
|
|
|
|
| 10 |
from dotenv import load_dotenv
|
| 11 |
|
| 12 |
from modules.answerer import answer_query
|
| 13 |
+
from modules.collector import fetch_channel_videos_by_id
|
| 14 |
from modules.db import get_collection
|
| 15 |
from modules.indexer import index_videos
|
| 16 |
|
|
|
|
| 26 |
|
| 27 |
# Collect + Index
|
| 28 |
for ch in CHANNELS:
|
| 29 |
+
videos = fetch_channel_videos_by_id(YT_API_KEY, ch)
|
| 30 |
index_videos(videos, collection)
|
| 31 |
|
| 32 |
# Ask a question
|
modules/collector.py
CHANGED
|
@@ -1,69 +1,65 @@
|
|
| 1 |
# -------------------------------
|
| 2 |
# 1. Collector
|
| 3 |
# -------------------------------
|
| 4 |
-
from typing import List,Dict
|
| 5 |
from googleapiclient.discovery import build
|
| 6 |
|
| 7 |
from modules.youtube_utils import get_channel_id
|
| 8 |
|
|
|
|
|
|
|
| 9 |
|
| 10 |
-
def
|
| 11 |
youtube = build("youtube", "v3", developerKey=api_key)
|
| 12 |
channel_id = get_channel_id(youtube, channel_url)
|
| 13 |
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
|
|
|
|
|
|
| 20 |
|
| 21 |
-
|
| 22 |
-
part="snippet",
|
| 23 |
-
channelId=channel_id,
|
| 24 |
-
maxResults=max_results,
|
| 25 |
-
order="date"
|
| 26 |
-
)
|
| 27 |
-
response = request.execute()
|
| 28 |
-
|
| 29 |
-
videos = []
|
| 30 |
-
for item in response.get("items", []):
|
| 31 |
-
if item["id"]["kind"] == "youtube#video":
|
| 32 |
-
videos.append({
|
| 33 |
-
"video_id": item["id"]["videoId"],
|
| 34 |
-
"title": item["snippet"]["title"],
|
| 35 |
-
"description": item["snippet"].get("description", ""),
|
| 36 |
-
"channel_id": channel_id,
|
| 37 |
-
"channel_title": channel_title,
|
| 38 |
-
})
|
| 39 |
-
return videos
|
| 40 |
-
|
| 41 |
-
def fetch_channel_videos(api_key: str, channel_id: str, max_results=20):
|
| 42 |
youtube = build("youtube", "v3", developerKey=api_key)
|
| 43 |
|
| 44 |
-
#
|
| 45 |
channel_response = youtube.channels().list(
|
| 46 |
-
part="snippet",
|
| 47 |
-
id=channel_id
|
| 48 |
).execute()
|
|
|
|
| 49 |
channel_title = channel_response["items"][0]["snippet"]["title"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 50 |
|
| 51 |
-
request = youtube.search().list(
|
| 52 |
-
part="snippet",
|
| 53 |
-
channelId=channel_id,
|
| 54 |
-
maxResults=max_results,
|
| 55 |
-
order="date"
|
| 56 |
-
)
|
| 57 |
-
response = request.execute()
|
| 58 |
-
|
| 59 |
-
videos = []
|
| 60 |
-
for item in response.get("items", []):
|
| 61 |
-
if item["id"]["kind"] == "youtube#video":
|
| 62 |
-
videos.append({
|
| 63 |
-
"video_id": item["id"]["videoId"],
|
| 64 |
-
"title": item["snippet"]["title"],
|
| 65 |
-
"description": item["snippet"].get("description", ""),
|
| 66 |
-
"channel_id": channel_id,
|
| 67 |
-
"channel_title": channel_title,
|
| 68 |
-
})
|
| 69 |
-
return videos
|
|
|
|
| 1 |
# -------------------------------
|
| 2 |
# 1. Collector
|
| 3 |
# -------------------------------
|
| 4 |
+
from typing import List, Dict
|
| 5 |
from googleapiclient.discovery import build
|
| 6 |
|
| 7 |
from modules.youtube_utils import get_channel_id
|
| 8 |
|
| 9 |
+
from googleapiclient.discovery import build
|
| 10 |
+
|
| 11 |
|
| 12 |
+
def fetch_all_channel_videos(api_key: str, channel_url: str, max_results_per_call=50):
|
| 13 |
youtube = build("youtube", "v3", developerKey=api_key)
|
| 14 |
channel_id = get_channel_id(youtube, channel_url)
|
| 15 |
|
| 16 |
+
final_videos = []
|
| 17 |
+
for videos in fetch_channel_videos_by_id(api_key, channel_id, max_results_per_call):
|
| 18 |
+
final_videos.extend(videos) # extend instead of append
|
| 19 |
+
print("Fetched", len(final_videos))
|
| 20 |
+
yield (f"Fetched {len(final_videos)}", final_videos)
|
| 21 |
+
|
| 22 |
+
yield (f"Fetched {len(final_videos)}", final_videos)
|
| 23 |
+
|
| 24 |
|
| 25 |
+
def fetch_channel_videos_by_id(api_key: str, channel_id: str, max_results=50):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
youtube = build("youtube", "v3", developerKey=api_key)
|
| 27 |
|
| 28 |
+
# Get uploads playlist ID
|
| 29 |
channel_response = youtube.channels().list(
|
| 30 |
+
part="contentDetails,snippet", id=channel_id
|
|
|
|
| 31 |
).execute()
|
| 32 |
+
|
| 33 |
channel_title = channel_response["items"][0]["snippet"]["title"]
|
| 34 |
+
uploads_playlist_id = channel_response["items"][0]["contentDetails"]["relatedPlaylists"]["uploads"]
|
| 35 |
+
|
| 36 |
+
next_page_token = None
|
| 37 |
+
|
| 38 |
+
while True:
|
| 39 |
+
request = youtube.playlistItems().list(
|
| 40 |
+
part="snippet",
|
| 41 |
+
playlistId=uploads_playlist_id,
|
| 42 |
+
maxResults=max_results,
|
| 43 |
+
pageToken=next_page_token,
|
| 44 |
+
)
|
| 45 |
+
response = request.execute()
|
| 46 |
+
|
| 47 |
+
videos = []
|
| 48 |
+
for item in response.get("items", []):
|
| 49 |
+
snippet = item["snippet"]
|
| 50 |
+
videos.append(
|
| 51 |
+
{
|
| 52 |
+
"video_id": snippet["resourceId"]["videoId"],
|
| 53 |
+
"title": snippet["title"],
|
| 54 |
+
"description": snippet.get("description", ""),
|
| 55 |
+
"channel_id": channel_id,
|
| 56 |
+
"channel_title": channel_title,
|
| 57 |
+
}
|
| 58 |
+
)
|
| 59 |
+
|
| 60 |
+
yield videos # yield one page worth
|
| 61 |
+
|
| 62 |
+
next_page_token = response.get("nextPageToken")
|
| 63 |
+
if not next_page_token:
|
| 64 |
+
break
|
| 65 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
modules/db.py
CHANGED
|
@@ -38,7 +38,7 @@ def get_indexed_channels(collection):
|
|
| 38 |
|
| 39 |
if cid: # only include if we have a channel_id
|
| 40 |
channels[cid] = cname
|
| 41 |
-
print("channels= ",channels)
|
| 42 |
return channels
|
| 43 |
|
| 44 |
|
|
@@ -48,8 +48,8 @@ def get_indexed_channels(collection):
|
|
| 48 |
def delete_channel_from_collection(channel_id: str):
|
| 49 |
"""Remove a channel from the index and refresh the radio choices."""
|
| 50 |
# Delete all videos for this channel
|
| 51 |
-
print("Deleting channel", channel_id)
|
| 52 |
data = get_collection().get(where={"channel_id": channel_id})
|
| 53 |
-
print("data = ", data)
|
| 54 |
get_collection().delete(where={"channel_id": channel_id})
|
| 55 |
collection = get_collection()
|
|
|
|
| 38 |
|
| 39 |
if cid: # only include if we have a channel_id
|
| 40 |
channels[cid] = cname
|
| 41 |
+
# print("channels= ",channels)
|
| 42 |
return channels
|
| 43 |
|
| 44 |
|
|
|
|
| 48 |
def delete_channel_from_collection(channel_id: str):
|
| 49 |
"""Remove a channel from the index and refresh the radio choices."""
|
| 50 |
# Delete all videos for this channel
|
| 51 |
+
# print("Deleting channel", channel_id)
|
| 52 |
data = get_collection().get(where={"channel_id": channel_id})
|
| 53 |
+
# print("data = ", data)
|
| 54 |
get_collection().delete(where={"channel_id": channel_id})
|
| 55 |
collection = get_collection()
|