scientes

> > > We also should rename the variables, using `a`, `b`, `c`, .. is bad practice and just makes it more difficult to understand/fix. yes sorry for that^^

the > > > What do you think of using [networkx](https://networkx.org/documentation/stable/reference/classes/digraph.html?highlight=directed%20graph) for the shortest path calculation? We could initiaize the graph with all pairs and calculate the shortest path from...

I'm not sure how to implement the config for exchanges. currently it is hardcoded in the config.py file, but ideally we should read the csvs in book and add them...

Also should i implement the networkx or should we go on with the custom graph.

I've tested a few times with own statements and fixed some bugs. But sadly i'm out of time for the next few months and wouldn't like this to go stale....

@provinzio what is still missing for a merge? I havent looked into this project for a while and wanted to get this merge finished

automatic I didn't implement because price_data is currently initialized before the book module which means that i build the graph before i know from which exchanges the csv's are from.

> Could it be that the resolution of the OHLCV is too low (e.g. only weekly data)? Shouldn't be. The weekly calls are only for looking for the time period...

> Hey @scientes, > > for the `ohlcv-batch` branch, I'm getting inaccurate price data for Kraken exports (see table below). The fetched prices are constant over multiple timestamps. I'm working...

https://www.freqtrade.io/en/stable/exchanges/#historic-kraken-data >The Kraken API does only provide 720 historic candles, which is sufficient for Freqtrade dry-run and live trade modes, but is a problem for backtesting. To download data for...