Download number of ratings and exact ratings from amazon (partly solved)

[Bücherwurm1]

There have been some requests for this feature, but I could not find a solution. So I created a "home-brewn" solution like this (unfortunately I am not familiar with python, so it is not integrated directly in Calibre):

1. Create two user-defined columns in Calibre. I named them "amazonrating" and "amazonnumratings".
2. Download meta data from Amazon so that an amazon-ID is there.
3. Filter to all books containing an Amazon-ID.
4. Create a catalogue with all these books. The catalogue should look like this:

id,identifiers
"373","isbn:9783548034522,amazon_de:3548034527,goo gle:IC8eSQAACAAJ"
"385","google:RnrIMgEACAAJ,isbn:9783841901705,amaz on_de:3841901700"
"388","amazon_de:3257201818,isbn:9783257201819 "
"541","google:LrsF5EE52OwC,amazon_de:3404921143,is bn:9783404921140

5. Run the attached (Source only) C#-program (requires (free) Microsoft Visual Studio Express 2013, .Net 4.5). The program fetches rating and number of ratings from the Amazon-Website based on the amazon-ID by doing some (probably not very reliable) parsing.
I would need some here: Can somebody provide me with the algorithm of the Amazon-plugin? I am sure this can be done much better than I did.

The output of the program is a windows batch-file writing the rating/number of ratings to calibre. It should look like this:

calibredb set_metadata 373 -f #amazonrating:5.0 -f #amazonnumratings:1
calibredb set_metadata 385 -f #amazonrating:4.0 -f #amazonnumratings:4
calibredb set_metadata 388 -f #amazonrating:4.8 -f #amazonnumratings:4
calibredb set_metadata 541 -f #amazonrating:4.5 -f #amazonnumratings:2

6. Start the batchfile.

This solution works find for me, but as a general solution there is one major drawback:
Paths, Websites, ..., are not configurable (they are hard-coded). Therefore e.g. this works for amazon.de only, and before you can use it, you have to change at least paths and compile on your own.

If some of you are interested, I can change that.

Some minor drawbacks:
1. Error handling could be improved
2. Parsing from the website could be improved
3. Performance (use more than one thread)
4. Probably more

Maybe this could help somebody,
Regards,
Bücherwurm1