import os
import time
import logging
import archivedb.logger
import archivedb.config as config
import archivedb.threads
import archivedb.sql
import archivedb.monitor
def _database_check():
    """ Perform preliminary database checks
    - Ensures required databases exist, and are created if they don't
    - Ensures required tables exist, and are created if they don't
    - Updates the watch_dir field based on what watch_dirs is set to in the conf
    @return: db
    @rtype: L{DatabaseConnection}
    ## perform db checks ##
    db_host = config.args["db_host"]
    db_port = config.args["db_port"]
    db_user = config.args["db_user"]
    db_pass = config.args["db_pass"]
    db_name = config.args["db_name"]
    db = archivedb.sql.DatabaseConnection(db_host, db_user, db_pass, db_name, db_port, "archive")
    # check if tables exist
    # add code to append to tables list if plugins are enabled
    tables = ["archive"]
    #c = db.cursor()
    for t in tables:
        if db.table_exists(t):
            log.debug("table '{0}' exists.".format(t))
  "required table '{0}' doesn't exist, creating.".format(t))
    # perform enum checks here
    # check if watch dirs in conf file match watch dirs set in database
    # use set() because it doesn't care about order
    conf_watch_dirs = set(args["watch_dirs"])
    db_watch_dirs = set(db.get_enum())
    if conf_watch_dirs == db_watch_dirs:
        log.debug("watch_dirs in conf match watch_dirs in database, no need to update enum")
        log.debug("watch_dirs in conf don't match to watch_dirs in database, need to update enum")
        log.debug("conf_watch_dirs    = {0}".format(conf_watch_dirs))
        log.debug("db_watch_dirs    = {0}".format(db_watch_dirs))
        print("\nATTENTION: Watch directories have been changed:")
        print("Old watch dirs: {0}".format(db_watch_dirs))
        print("New watch dirs: {0}".format(conf_watch_dirs))
        print("If any watch directories have been removed, ",
              "they will be removed permanently from the database.")
        # py3 doesn't use raw_input(), and input() for py2 expects an int
        if config._py3: resp = input("Update? ").lower()
        else: resp = raw_input("Update? ").lower() #@UndefinedVariable
        if resp in ('y', 'yes'):
            db.alter_enum("watch_dir", args["watch_dirs"])
            print("OK, will not update watch directories in the database.")
def clean():
    log.debug("start: clean()")
    rows_cleaned = 0
    db = _database_check()
    # remove all rows where watch_dir = '' 
    # (this would occur when a watch_dir has been removed from the enum list)
    sql = "DELETE FROM `archive` WHERE watch_dir = ''"
    log.debug("executing: \"{0}\"".format(sql))
    rows = db._execute(sql)
    log.debug("{0} rows removed from last query.".format(rows))
    rows_cleaned += rows
    # check all files in the database for their existence, remove if not found
    ROW_COUNT = 500
    row_offset = 0
    del_ids = []"purging nonexistent files in the database (may take awhile)")
    while True:
        sql = """SELECT id, watch_dir, path, filename 
        FROM `archive` LIMIT {0}, {1}""".format(row_offset, ROW_COUNT)
        log.debug("executing: \"{0}\"".format(sql))
        rows = db._query(sql)
        for r in rows:
            full_path = os.path.join(*r[1:])
            if not os.path.exists(full_path):
                log.debug("marked for deletion: {0}".format(full_path))
        # if the number of rows received from the query is equal to ROW_COUNT,
        # that means there's probably more rows left to check. But if it's
        # less than ROW_COUNT, that means we've reached the end of the database
        if len(rows) < ROW_COUNT: break
        # update the offset
        else: row_offset += len(rows)
    for i in del_ids:
        rows_cleaned += 1"total rows cleaned: {0}".format(rows_cleaned))
    log.debug("end: clean()")
def main():
    ## create threads ##
    threads_dict = archivedb.threads.initialize_child_processes(args["threads"])
    while True:
        threads_dict = archivedb.threads.keep_child_processes_alive(threads_dict)
if __name__ == 'archivedb.core':
    args = config.args
    log = logging.getLogger(__name__)