#!/usr/bin/env python


import datetime
import logging
import json
import os
import subprocess
import sys
import yaml


CREOLE_JAR = "/home/eevans/creole-cli-1.0.0-SNAPSHOT-jar-with-dependencies.jar"
SSUTIL_JAR = "/home/eevans/twcs-inspector-1.0.0-SNAPSHOT.jar"


os.environ["FORMAT"] = "json"
logging.basicConfig(level=logging.DEBUG)


def discover_instance_jmx(directory):
    for name in os.listdir("/etc/cassandra-instances.d"):
        if not name.endswith(".yaml"):
            continue
        with open(os.path.join("/etc/cassandra-instances.d", name)) as f:
            y = yaml.load(f)
            for data_directory in y["data_file_directories"]:
                if directory.startswith(data_directory):
                    return y["jmx_port"]
    return -1

def setup(directory):
    # TODO: validate jars are present
    # TODO: validate existence, readability, of directory supplied
    # TODO: lock, so that we do not trample upon another long-running instance
    pass

def main():
    if len(sys.argv) != 2:
        print >>sys.stderr, "Usage: {} <directory>".format(sys.argv[0])
        sys.exit(1)

    directory = sys.argv[1]
    setup(directory)
    __ms_to_date = lambda ms: datetime.datetime.fromtimestamp(int(ms) / 1e3).strftime("%Y-%m-%d")

    tables = json.loads(subprocess.check_output(["java", "-jar", SSUTIL_JAR, directory]))
    if not len(tables) > 2:
        logging.warn("too few windows to begin a new compaction")
        return 1
    windows = sorted(tables.keys())
    candidates = []
    data_files = []
    for candidate in windows[:len(windows)-1]:
        # This should be all but the most current window, each of which should have
        # been compacted down to a single table.  If that is not the case (perhaps),
        # the compaction is on-going, then exclude the window for this iteration.
        if len(tables[candidate]) > 1:
            logging.warn(
                "window starting %s hasn't been fully compacted, stopping here...",
                __ms_to_date(candidate))
            # Do not consider any later windows, either...
            break
        candidates.append(candidate)
        table = tables[candidate][0]
        logging.debug(
            "adding window starting %s, data file %s (%s percent droppable)",
            __ms_to_date(candidate),
            table["dataFileName"],
            table["estimatedDroppableTombstones"] * 100)
        data_files.append(os.path.join(directory, table["dataFileName"]))

    if len(data_files) < 2:
        logging.warn("not enough candidates for compaction, exiting...")
        return 0

    jmx_port = discover_instance_jmx(directory)
    if jmx_port < 0:
        logging.error("Unable to detect instance from directory")
        return 1
    subprocess.call(["java", "-jar", CREOLE_JAR, "-H", "localhost", "-P", str(jmx_port), "user-compaction"] + data_files)
    return 0


if __name__ == "__main__":
    sys.exit(main())