nullhypothesis / sybilhunter Goto Github PK

Sybilhunter's churn module should also be able to determine the fraction of consensus weight joining or leaving the network, like NewCW and GoneCW.

Thanks to Karsten for suggesting this:
https://lists.torproject.org/pipermail/metrics-team/2016-February/000069.html

Running:

$ go get github.com/NullHypothesis/sybilhunter

Returns this to the console:

# github.com/NullHypothesis/sybilhunter
../../NullHypothesis/sybilhunter/uptime.go:90: constant 4278190080 overflows int

Program will not compile, on a 32-bit machine with 3MB of ram.

Currently, sybilhunter uses strict parsing to pass objects to modules. Some modules, however, only need relay fingerprints, which means that lazy parsing is sufficient. Modules should be able to express that they prefer lazy parsing.

We want to be able to extract all consensuses and descriptors that contain a relay identified by its IP address or by its fingerprint. Ideally, we also want to specify a data range, so we don't have to go through the entire archives.

I'd like to generate uptime images only about relays with a given nickname or a given nickname prefix.

The architecture should consist of a file processor that parses all given files and calls the analysis functions (in a goroutine) as soon as parsing is done. That makes it easy to run several analysis functions at once.

I am new to go, can you help me with this?

$ go get -v github.com/NullHypothesis/sybilhunter
git.torproject.org/user/phw/zoossh.git (download)
github.com/DataWraith/vptree (download)
github.com/NullHypothesis/mlgo (download)
github.com/arbovm/levenshtein (download)
github.com/mcgrew/gostats (download)
github.com/NullHypothesis/mlgo
github.com/arbovm/levenshtein
github.com/DataWraith/vptree
github.com/NullHypothesis/mlgo/cluster
github.com/mcgrew/gostats
git.torproject.org/user/phw/zoossh.git
github.com/NullHypothesis/sybilhunter

$ sybilhunter
sybilhunter: command not found

Instead of analysing every single consensus, we might want to analyse only one per week or per month.

Thanks to Karsten for suggesting this:
https://lists.torproject.org/pipermail/metrics-team/2016-February/000069.html

Thanks to Karsten for suggesting this:
https://lists.torproject.org/pipermail/metrics-team/2016-February/000069.html