voidstar0 / blm-instagram-bot

Inform users clogging the #BlackLivesMatter tag on IG with solid images

blm-instagram-bot

Instagram bot that informs users using the BlackLivesMatter tag to post solid images that they may be accidentally hiding important information about the movement.

Uses a Google Cloud Function called blm-cloud-function to determine if the image is a solid black image. This is to offload the image processing to the cloud and hopefully keep things running smooth.

Setup

You need Python 3 and pip. On Ubuntu: sudo apt install python3-dev python3-pip will get you there. You will use python3 and pip3 in place of pip and python, unless you override which the default is with your $PATH variable.

1. Create a burner account or two. Edit accounts.json to include their login information.
2. Edit the bios to include relevant information. Example
3. Run pip install -r requirements.txt. If you are not on Windows, comment out the lines with pywin32 and pypiwin32.
4. Rename .env.example to .env
5. Run python3 bot.py. If you get a missing module error, python3 -m pip install -r requirements.txt and try again.

To-Do

1. Poll the tag. Right now it only goes through the current posts and stops.

Make sure to rename .env.example to .env Also make sure you're using Python 3.5 or greater.

blm-instagram-bot's Issues

Enhancement: Scaling with Celery

I saw that you mentioned looking for assistance with scaling. Using a distributed computing library to assign work to a group of "worker" nodes would be a nice way to implement this. I'm familiar with a Python library called Celery that does just that. The main question is how to set up a release distribution that enables individuals to spin up their own workers easily on whatever computing resources they have available.

Additionally, I imagine querying for and detecting image matches is the most computationally expensive task in this project. With this in mind, a distributed system will require a way to designate "chunks of work" to individual instances/nodes so that repeat queries are avoided where possible.

I'll look deeper into the code to see if this can be easily accomplished with the current setup.

Rate Limited

Not sure if this issue is solely with my implementation but for some reason I am constantly being given the "rate limited" error message, along with an index error from the line of code "post = feed['items'].pop(0).

Rename the master branch to main

The term "master" has racist and negative connotations leading back to slavery, which undermines the efforts of this repository. (See: https://www.bbc.com/news/technology-53050955)

Because of this I believe that we should rename the master branch to main, so that we may continue to make black people feel comfortable in this community.

Use smallest image from image_versions2

Currently the bot is just picking out the zeroth element:

url = post['image_versions2']['candidates'][0]['url']

But it would be more efficient if we sorted by size and picked the smallest image to send to the detection API.

I've done this in JS (am working on a node version bc I don't know python) but in the meantime if someone wants to port this to python I think this would be an improvement:

const smallestImage = item.image_versions2.candidates.sort((a, b) => a.width * a.height > b.width * b.height).pop();

#BlackoutTuesday getting merged into #BlackLivesMatter?

I've heard some anecdotal reports that Instagram is merging the two hashtags. Does it make sense to change the script of the comment to urge people to post links to local bail funds, mutual aid organizations, protest groups, NAACP chapters, etc?

TODO: Detection method for black photos

Currently the logic checks if all pixels in all channels are below 20.

How well is this working? If it's fine, keep it.

If you want to do something else, here are some suggestions based on my experience with image processing:

Take the average pixel value across all channels. Threshold it to <5. That way, even small blips of white that may be there will get washed out in the process of averaging. From what I understand, having a single pixel of (255,255,255) right now would prevent the logic from returning True (I'm not native to JS but that's my understanding from the cloudfunction repo's index.js).

Take the median. Same logic.

Sum all three channels, divide by 3. If the proportion of pixels below (5,5,5) exceeds 95% / 99%, mark it as solid.

JPEG compression really in theory shouldn't be leaving artifacts in all-black images. But if someone clips just a bit of white in there, we want to be able to catch that.

change default message

current message assumes info in bio. needs to be more general, point to a specific account (blm_tag_bot)