implement feedback corrections for the search based on the previous development: it searches for old posts a lot, and duplicates appear. I commented in the report for September 20 https://docs.google.com/spreadsheets/d/15mCC3ssEGTU-HWM0GR1eG0sXE9zUtVqNQJCKBdWxTaQ/edit?gid=0#gid=0

add the date to the advanced search - 2 hours

Check the number of subscribers, because there is a filter, but you noticed that the check does not work, the code is commented out.

requires preliminary analysis - 2-4 hours

Captcha - handling cases when a Captcha appears, and solving the Captcha.

First stage, detection and identification - 4 hours (based on the results we will make a decision regarding the feasibility of solving it, I remind you that manual solving during tests did not yield the desired result)

Add an option for logging into Twitter accounts that the bot works with using login and password, so that the bot automatically retrieves cookies and then automatically retrieves keys from it and enters them into the table. Currently, the bot works with accounts using tokens, but there are suspicions that tokens may have a certain lifespan of 30-60 days, or until Twitter may suspect the account of suspicious activity (this can happen even with accounts that do not violate anything), if a code is requested from the email linked to the account, then make a solution for obtaining this code via API, accounts will mainly be on rambler, @gmx.com and Google;

research the possibility of accessing email from the script in the domain gmx.com, writing a test script (from you, if there are account credentials) - 4 hours
if the previous point is successful, writing general code for account authorization, a general part for working with Twitter, and separate modules for working with different email servers (to have the ability to change or add another) - 20 - 30 hours --- this is for the next project

Accounts for shilling and warming up should be taken randomly, not in order. Because the first accounts bear the load and they get killed quickly.

we make a random selection - 2 hours

Check if this functionality is needed:

Additionally, we are looking for "Main words" in the post text (field 1), I left this as it is for now. If there is no need for this functionality, then they can be omitted. Or this functionality can be removed.

I will test and most likely remove it - 2 hours

handling cases when posts are not found

if we finish the cycle, we will take a break and try again - 1 hour

if you have another logic, like notifying the user that there are no posts for this request, for example after several attempts, then write, we will discuss. Because now that the filters are working, it can easily happen that there are no posts for these requests and this is not very obvious.

When you fill out the project and click outside the project window, it closes. Very inconvenient, often you have to start filling out from the beginning. It needs to be made so that the project window only closes with the close button.

2 hours

Additionally, during development, code refactoring (removing unnecessary parts, fixing what I see)

Proposals 5