analyze.jl

using Pandas
using PyPlot

function to_df()
    issues = deserialize(open("issues.jld"))
    df = DataFrame(issues)
    for datecol in ["closed_at", "created_at"]
        df[datecol] = to_datetime(df[datecol])
    end
    to_pickle(df, "issues.pickle")
    df
end

to_inc(s) = Series(ones(length(s)), index=s)

function analyze_ts(df)
    create = to_inc(df["created_at"])
    figure()
    hold(true)
    plot_by_week(create)
    close = to_inc(dropna(df["closed_at"]))
    plot_by_week(close)
    legend(["Created", "Closed"])
    ylabel("# of issues")
    xlabel("Time")
    savefig("issues.pdf")
end

plot_by_week(s) = plot(resample(s, "1W", how="sum"))

issues.pdf

issues.png

scrape_issues.jl

using Requests
using JSON

auth_token = readall(open("auth_token.txt"))

N_issues = [645, 5179]
issues = {}

for (N_issue, state) in zip(N_issues, ["open", "closed"])
    n_requests = int(ceil(N_issue/100))
    for i=1:n_requests
        @show i
        s = string("https://api.github.com/repos/JuliaLang/Julia/issues?access_token=", auth_token, "&page=", i, "&per_page=100", "&state=", state)
        local r
        while true
            try
                r = get(s)
                break
            catch err
                warn(err)
            end
        end
        j = JSON.parse(r.data)
        issues = [issues; j]
        @show length(issues)
    end
end

info("writing")
f=open("issues.jld","w")
serialize(f, issues)