Obsei: Missing YouTube dislike count?

Lalit Pagaria — Thu, 10 Feb 2022 08:29:50 +0000

Dear Dev Community!

Obsei is an open-source low-code AI powered automation tool. Please refer Obsei introduction thread. I would like to share about our new integration: YouTube video comments.

As you are aware that YouTube is stopped showing video "dislike" count. So how about analyzing it via sentiments of the comments?

With this integration you can easily perform following tasks:

getting alert on negative or spam comment on your video
aggregate all comments and extract insight to analyze reveal more detailed engagement metrics
fast and easy creation of ML review dataset
and many more

Please check the demo:

If you are Python person then this is all you need it:

from obsei.source.youtube_scrapper import YoutubeScrapperSource, YoutubeScrapperConfig

# initialize Youtube source config
source_config = YoutubeScrapperConfig(
    video_url="https://www.youtube.com/watch?v=uZfns0JIlFk", # Youtube video URL
    fetch_replies=True, # Fetch replies to comments
    max_comments=10, # Total number of comments and replies to fetch
    lookup_period="1Y", # Lookup period from current time, format: `<number><d|h|m|M|Y>` (day|hour|minute|month|year)
)

# initialize Youtube comments retriever
source = YoutubeScrapperSource()

# This will fetch information from configured source ie twitter, app store etc
source_response_list = source.lookup(source_config)

If like YAML then you would need following config:

source:
      _target_: obsei.source.youtube_scrapper.YoutubeScrapperSource
source_config:
      _target_: obsei.source.youtube_scrapper.YoutubeScrapperConfig
      video_url: "https://www.youtube.com/watch?v=uZfns0JIlFk"
      lookup_period: "1Y"
      max_comments: 10

Looking forward for your valuable feedback.

Repo URL: https://github.com/obsei/obsei

Introducing Obsei

Lalit Pagaria — Thu, 13 Jan 2022 11:03:03 +0000

Dear Dev Community!

Allow me to introduce Obsei (pronounced "Ob see" | /əb-'sē/). It is a low-code AI powered automation tool. Name was derived from three words OBServe, AnalyzE, Inform.

You can think of it as Zapier for cognitive-automation tasks. More specifically we closely resemble Levity, which are doing great work toward no-code ML automation.

Following GIF can give you an overview of Obsei's -

Features

Apache 2.0 License: Permissible FOSS license
Low-Code: YAML is all you need
Stateless: Privacy and serverless friendly
GitHub Action: Schedule workflow via GitHub Actions (This might save you Zaps for few tasks)
BYOM: Bring your own model (via HuggingFace Hub)
Multi-Mode: Currently support text based workflow but future plan to support Image, Audio, OCR and Video as well

Our main idea is to provide less complex Lego pieces, which can help less tech savvy people to build complex workflows with ease.

Currently 4 companies are using Obsei for their use cases. Superblog shared how Obsei help them to find more qualified sales leads on Twitter (filtering relevant information via AI). Tweet link

We are bootstrapped small startup, looking forward for community support and feedback to improve and making Obsei more useful. Code is still in early stage and next 2 months we will release more stable Beta version along with properly taking care of backward compatibility and breaking changes.

Refer important links:
Repo: https://github.com/obsei/obsei
Walkthrough Video: Youtube link
Demo UI: https://huggingface.co/spaces/obsei/obsei-demo
Demo UI video: Youtube link

DEV Community: Lalit Pagaria

Obsei: Missing YouTube dislike count?

Introducing Obsei

Features