cortesi / devsurvey

The dataset is a PostgreSQL dump file, compressed with 7zip.
Finding the primary language for each project was tricky, and the results are
inexact. Language detection was done purely on the basis of file extension. The
map between the strings in the "primary" field and languages is as follows:
    ActionScript:       "as"
    C++:                "cpp"
    C:                  "c"
    Clojure:            "clj"
    C#:                 "cs"
    Erlang:             "erl"
    Go:                 "go"
    Haskell:            "hs"
    Java:               "java"
    Javscript:          "js"
    Lisp:               "lisp"
    Lua:                "lua"
    Objective C:        "m"
    Objective J:        "j"
    PHP:                "php"
    Ruby:               "rb"
    Perl:               "pl"
    Python:             "py"
    Scala:              "scala"
    
The schema is as follows:
CREATE TABLE author_repos (
    id integer NOT NULL,
    repo_id integer NOT NULL,
    author_id integer NOT NULL,
    count integer                                       -- Total # commits
);
CREATE TABLE authors (
    id integer NOT NULL,
    name character varying(100) NOT NULL                -- Author email
);
CREATE TABLE commits (
    id integer NOT NULL,
    repo_id integer NOT NULL,
    author_id integer NOT NULL,
    age integer NOT NULL,                               -- Commit age in seconds, with 0 being the time of the first 
    commit to the project
    message text,
    deletions integer,                                  -- lines deleted
    insertions integer,                                 -- lines inserted
    lines integer,                                      -- deletions + insertions
    files integer,                                      -- Number of files touched
    hash character varying(42)                          -- SHA hash
);
CREATE TABLE extensions (
    id integer NOT NULL,
    repo_id integer NOT NULL,
    ext character varying(6) NOT NULL,                  -- file extension
    count integer                                       -- total number of lines
);
CREATE TABLE repos (
    id integer NOT NULL,
    "user" character varying(100) NOT NULL,
    name character varying(100) NOT NULL,
    watchers integer,                                   -- github watchers
    forks integer,                                      -- github forks
    "primary" character varying(6),                     -- primary language, calculated from extensions
    start integer,                                      -- first commit, seconds since epoch
    "end" integer,                                      -- last commit, seconds since epoch
    commitcount integer,                                -- total commits
    authorcount integer,                                -- total authors
    insertions integer,                                 -- total lines inserted 
    deletions integer,                                  -- total lines deleted
    lines integer                                       -- insertions + deletions
);