pferd/PFERD/__main__.py

127 lines
3.8 KiB
Python
Raw Normal View History

2021-05-13 23:52:46 +02:00
import argparse
import asyncio
2021-05-15 21:33:51 +02:00
import configparser
2021-05-13 23:52:46 +02:00
from pathlib import Path
from .cli import PARSER, load_default_section
from .config import Config, ConfigDumpError, ConfigLoadError, ConfigOptionError
from .logging import log
from .pferd import Pferd, PferdLoadError
2021-05-22 22:38:56 +02:00
from .transformer import RuleParseError
2021-05-19 17:32:23 +02:00
from .version import NAME, VERSION
2021-05-13 23:52:46 +02:00
2021-05-15 21:33:51 +02:00
def load_config_parser(args: argparse.Namespace) -> configparser.ConfigParser:
2021-05-19 18:10:17 +02:00
log.explain_topic("Loading config")
2021-05-15 21:33:51 +02:00
parser = configparser.ConfigParser()
if args.command is None:
2021-05-19 18:10:17 +02:00
log.explain("No CLI command specified, loading config from file")
2021-05-15 21:33:51 +02:00
Config.load_parser(parser, path=args.config)
2021-05-19 18:10:17 +02:00
else:
log.explain(f"CLI command specified, creating config for {args.command!r}")
if args.command:
args.command(args, parser)
2021-05-15 21:33:51 +02:00
load_default_section(args, parser)
2021-05-15 21:33:51 +02:00
return parser
def load_config(args: argparse.Namespace) -> Config:
try:
return Config(load_config_parser(args))
except ConfigLoadError as e:
log.error(str(e))
log.error_contd(e.reason)
exit(1)
2021-05-13 23:52:46 +02:00
def configure_logging_from_args(args: argparse.Namespace) -> None:
if args.explain is not None:
log.output_explain = args.explain
# We want to prevent any unnecessary output if we're printing the config to
# stdout, otherwise it would not be a valid config file.
if args.dump_config == "-":
2021-05-19 18:10:17 +02:00
log.output_explain = False
def configure_logging_from_config(args: argparse.Namespace, config: Config) -> None:
# In configure_logging_from_args(), all normal logging is already disabled
# whenever we dump the config. We don't want to override that decision with
# values from the config file.
if args.dump_config == "-":
return
try:
if args.explain is None:
log.output_explain = config.default_section.explain()
except ConfigOptionError as e:
log.error(str(e))
exit(1)
def dump_config(args: argparse.Namespace, config: Config) -> None:
2021-05-23 11:04:50 +02:00
log.explain_topic("Dumping config")
try:
if args.dump_config is True:
config.dump()
elif args.dump_config == "-":
config.dump_to_stdout()
else:
config.dump(Path(args.dump_config))
except ConfigDumpError as e:
log.error(str(e))
log.error_contd(e.reason)
exit(1)
def main() -> None:
args = PARSER.parse_args()
2021-05-19 17:32:23 +02:00
if args.version:
print(f"{NAME} {VERSION}")
exit()
# Configuring logging happens in two stages because CLI args have
# precedence over config file options and loading the config already
# produces some kinds of log messages (usually only explain()-s).
configure_logging_from_args(args)
2021-05-13 23:52:46 +02:00
config = load_config(args)
# Now, after loading the config file, we can apply its logging settings in
# all places that were not already covered by CLI args.
configure_logging_from_config(args, config)
2021-05-15 21:33:51 +02:00
if args.dump_config is not None:
dump_config(args, config)
2021-05-13 23:52:46 +02:00
exit()
2021-05-22 21:05:32 +02:00
try:
pferd = Pferd(config, args.crawler)
asyncio.run(pferd.run())
except (PferdLoadError, ConfigOptionError) as e:
log.unlock()
2021-05-22 21:05:32 +02:00
log.error(str(e))
exit(1)
2021-05-22 22:38:56 +02:00
except RuleParseError as e:
log.unlock()
e.pretty_print()
exit(1)
2021-05-22 14:45:32 +02:00
except KeyboardInterrupt:
log.unlock()
log.explain_topic("Interrupted, exiting immediately")
log.explain("Open files and connections are left for the OS to clean up")
log.explain("Temporary files are not cleaned up")
2021-05-22 14:45:32 +02:00
# TODO Clean up tmp files
# And when those files *do* actually get cleaned up properly,
# reconsider what exit code to use here.
exit(1)
except Exception:
log.unlock()
log.unexpected_exception()
exit(1)