TagSearcher is a tool that fetches all the posts made for a given tag between the given time period. You can access this tool by heading over to https://scribe.steemcryptic.me/tagsearcher

^{New version of Scribe}

Hey all. I know I haven't been very active lately. This was because of the huge amount of work due to a release at the office. The release is scheduled to happen today. So, I will be more active over here after the release. I did, however, find time to improve TagSearcher to make it more accurate.

Problem:

TagSearcher was a tool that I had developed mainly to help the crypto academy professors rate the homework assignments in a more organized manner. Looking at its usage every day, I feel that it has been able to achieve this goal up to some extent.

TagSearcher made use of the active votes property of every academy post to determine if it was visited by the professors and the steemcurators. Last weekend, I noticed two main problems with this design.
1- There are times where a few posts are auto-voted using steemauto. In reality, the professors would not have visited the posts yet.
2- There have also been numerous incidents wherein the professors just rate a post and don't vote it.

In both these scenarios, the information fetched by TagSearcher ends up being incorrect.

Solution:

<sup>Incorrect data before the fix



After the fix, row 10 shows the right data</sup>

The obvious solution was to also fetch the comments for the particular post and determine if a professor had visited the post based on that. However, trying to fetch the comments meant that I would need to make an extra API call to the RPC nodes which in turn meant more time to retrieve the results.

This is where multi-threading came in very handy. By making all the calls to get_content_comments API in an asynchronous fashion, I was able to achieve a performance that was very close to the previous performance.

Apart from the major change to the design of TagSearcher, I also fixed a bug that was introduced last week. This was related to a few posts not getting picked by TagSearcher.

Future Enhancements:

The performance of TagSearcher isn't as great as I would want it to be. There are several synchronous API calls that happen. So, my next goal is to make all the calls asynchronous to make the whole processing a lot faster. Since this would mean a complete redesign, I am in a dilemma whether to wait for @steemchiller's SDS implementation of this or just spend some time redesigning.

Apart from that, the tag searcher is also missing the total account SP details which would require another API call. Once these two features are implemented, TagSearcher should be a whole lot more useful.

I am also open to extending TagSearcher support to other communities or Use cases. If you have any request for the same, drop a comment down below and I will get to it ASAP.

I would like to thank @sapwood, @steemcurator01, @steemitblog as well as everyone who uses tag searcher for showing such great support for the tool

Thank you.

CC:
@sapwood
@steemcurator01
@steemitblog
@steemcurator02