#!/usr/bin/python

import sqlalchemy

from dedup.utils import fetchiter

def add_values(conn, insert_key, files, size):
    rows = conn.execute("UPDATE sharing SET files = files + ?, size = size + ? WHERE package1 = ? AND package2 = ? AND func1 = ? AND func2 = ?;",
                        (files, size) + insert_key)
    if rows.rowcount > 0:
        return
    conn.execute("INSERT INTO sharing (package1, package2, func1, func2, files, size) VALUES (?, ?, ?, ?, ?, ?);",
                 insert_key + (files, size))

def compute_pkgdict(rows):
    pkgdict = dict()
    for package, _, filename, size, function in rows:
        funcdict = pkgdict.setdefault(package, {})
        funcdict.setdefault(function, []).append((size, filename))
    return pkgdict

def process_pkgdict(conn, pkgdict):
    for package1, funcdict1 in pkgdict.items():
        for function1, files in funcdict1.items():
            numfiles = len(files)
            size = sum(entry[0] for entry in files)
            for package2, funcdict2 in pkgdict.items():
                if package1 == package2:
                    pkgnumfiles = numfiles - 1
                    pkgsize = size - min(entry[0] for entry in files)
                    if pkgnumfiles == 0:
                        continue
                else:
                    pkgnumfiles = numfiles
                    pkgsize = size
                for function2 in funcdict2.keys():
                    insert_key = (package1, package2, function1, function2)
                    add_values(conn, insert_key, pkgnumfiles, pkgsize)

def main():
    db = sqlalchemy.create_engine("sqlite:///test.sqlite3")
    with db.begin() as conn:
        conn.execute("PRAGMA foreign_keys = ON;")
        conn.execute("DELETE FROM sharing;")
        conn.execute("DELETE FROM duplicate;")
        readcur = conn.execute("SELECT hash FROM hash GROUP BY hash HAVING count(*) > 1;")
        for hashvalue, in fetchiter(readcur):
            rows = conn.execute("SELECT content.package, content.id, content.filename, content.size, hash.function FROM hash JOIN content ON hash.cid = content.id WHERE hash = ?;",
                                (hashvalue,)).fetchall()
            print("processing hash %s with %d entries" % (hashvalue, len(rows)))
            pkgdict = compute_pkgdict(rows)
            conn.execute("INSERT OR IGNORE INTO duplicate (cid) VALUES (?);",
                         *[(row[1],) for row in rows])
            process_pkgdict(conn, pkgdict)

if __name__ == "__main__":
    main()