Mat Kelly

Ok. Per above, some code snippets, timing commands, and sample data would make this more empirically convincing as a better solution than what is currently implemented in ipwb.

https://github.com/machawk1/cdxjGenerator/

I created a CDXJ file using the code in the aforementioned repo with 500,000 entries and tried replaying it in ipwb via `ipwb replay sample_sorted.cdxj`. Loading up `localhost:5000` in the...

@ibnesayeed 300+MB (with a spike) difference handling the request. For the sake of Counting Mementos™, there might be a more memory efficient, quicker (

On a 500,000 entry CDXJ file, `time curl http://localhost:5000/ > /dev/null` requires 1m35s to resolve for me using latest master 63433f3 and `ipwb replay sample.cdxj`. Related aside, `getURIsAndDatetimesInCDXJ()` is called...

I extracted the relevant parts from the code to which @ibnesayeed linked from MementoMap and appended them below to be integrated into ipwb. This executes and returns the value in...

issue-604 branch uses the aforementioned binary search algorithm and is effective at displaying the initial UI with very large CDXJ files in a reasonable amount of time. As above, the...

@anatoly-scherbakov Reiterating @ibnesayeed, thanks for your interest. While I am open to using other data sources for K/V lookup, one goal is for the CDXJ indexes to be shareable with...

> I would prefer keeping the replay and indexer separate from the presentation layer. That's the idea. What about something akin to memgator's web interface? Regarding rewriting, thoughts on using...

Though it's not what I think I intended with this ticket (that being a standalone-ish app), the replay system is aware of the state of the daemon and has toggle...