ArchiveBox/archivebox/extractors/extractor.py

import hashlib
import mimetypes
import os

import subprocess
from typing import ClassVar
from datetime import timedelta
from zipfile import Path

from django.utils import timezone

from archivebox.misc.hashing import get_dir_info

from core.models import ArchiveResult

import abx
import archivebox

class Extractor:
    # static class variables
    name: ClassVar[str] = 'ytdlp'
    verbose_name: ClassVar[str] = 'YT-DLP'
    binaries: ClassVar[tuple[str, ...]] = ()
    daemons: ClassVar[tuple[str, ...]] = ()
    timeout: ClassVar[int] = 60
    
    # instance variables
    ARCHIVERESULT: ArchiveResult
    CONFIG: dict[str, object]
    BINARIES: dict[str, object]
    DAEMONS: dict[str, object]
    
    def __init__(self, archiveresult: ArchiveResult, extra_config: dict | None=None):
        assert archiveresult.pk, 'ArchiveResult must be saved to DB before it can be extracted'
        self.archiveresult = self.ARCHIVERESULT = archiveresult
        self.CONFIG = archivebox.pm.hook.get_SCOPE_CONFIG(archiveresult=self.archiveresult, extra=extra_config)
        all_binaries = abx.as_dict(archivebox.pm.hook.get_BINARIES())
        all_daemons = abx.as_dict(archivebox.pm.hook.get_DAEMONS())
        self.BINARIES = {
            binary_name: all_binaries[binary_name]
            for binary_name in self.binaries
        }
        self.DAEMONS = {
            daemon_name: all_daemons[daemon_name]
            for daemon_name in self.daemons
        }

    def extract(self, config: dict | None=None) -> 'ArchiveResult':
        """
        - making sure any binaries the extractor depends on are installed and loaded
        - creating a new temporary working directory under the snapshot dir to hold extractor output
        - setting up a timer signal to kill the extractor if it runs too long
        - passing the extractor the URLs, temporary working directory, and config dict of options
        - running the extractor in a shell subprocess and collecting stdout/stderr
        - capturing the extractor's exit code
        - if extractor exits with 29 (RetryError), it should set the status to 'BACKOFF' and set retry_at to a datetime in the future
        - if extractor exits with 50 (NotApplicable), it should set the status to 'SKIPPED', and set retry_at to None
        - setting the correct permissions and ownership on all the output files
        - generating the merkle tree of all the output files and their hashes
        - generating a thumbnail of the main output (or collecting one provided by the extractor)
        - detecting any special outputs files that need to be parsed for other parts of the system (content-types? )
            - metadata.json -> ArchiveResult.output_json
            - outlinks.jsonl -> ArchiveResult.output_links
            - search_texts.txt -> ArchiveResult.index_texts
            - .merkle.json -> ArchiveResult.output_files
            - videos.jsonl -> ArchiveResult.output_videos
            - audios.jsonl -> ArchiveResult.output_audios
            - images.jsonl -> ArchiveResult.output_images
            - htmls.jsonl -> ArchiveResult.output_htmls
        - saving all the result metadata to the ArchiveResult in the database
        """
        
        archiveresult = self.ARCHIVERESULT
        # config = get_scope_config(archiveresult=archiveresult.snapshot.url, env=...)
        
        self.before_extract()

        error = Exception('Failed to start extractor')
        stdout = ''
        stderr = ''
        try:
            proc = archiveresult.EXTRACTOR.spawn(url=archiveresult.snapshot.url, binaries=binaries, daemons=daemons, cwd=cwd, config=config)
            stdout, stderr = proc.communicate()
            error = None
        except Exception as err:
            error = err
        finally:
            self.after_extract(error=error)
        
        return archiveresult
        
    def should_extract(self):
        if self.archiveresult.snapshot.url.startswith('https://youtube.com/'):
            return True
        return False

    def load_binaries(self):
        return {
            bin_name: binary.load()
            for bin_name, binary in self.BINARIES.items()
        }
    
    def load_daemons(self):
        return {
            daemon_name: daemon.load()
            for daemon_name, daemon in self.DAEMONS.items()
        }
        
    def output_dir_name(self):
        # e.g. 'ytdlp'
        return f'{self.name}'
    
    @property
    def OUTPUT_DIR(self):
        return self.archiveresult.snapshot_dir / self.output_dir_name()
    
    def before_extract(self):
        # create self.archiveresult.snapshot_dir / self.archiveresult.extractor / dir
        # chown, chmod, etc.
        binaries = self.load_binaries()
        daemons = self.load_daemons()
        cmd = self.archiveresult.EXTRACTOR.get_cmd(binaries=binaries, daemons=daemons)
        cmd_version = self.archiveresult.EXTRACTOR.get_cmd_version(binaries=binaries, daemons=daemons)
        
        self.OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
        os.chmod(self.OUTPUT_DIR, 0o755)
        self.archiveresult.status = self.archiveresult.StatusChoices.STARTED
        self.archiveresult.retry_at = timezone.now() + timedelta(seconds=self.timeout)
        self.archiveresult.start_ts = timezone.now()
        self.archiveresult.end_ts = None
        self.archiveresult.output = None
        self.archiveresult.output_path = str(self.OUTPUT_DIR.relative_to(self.archiveresult.snapshot_dir))
        self.archiveresult.cmd = cmd
        self.archiveresult.cmd_version = cmd_version
        self.archiveresult.machine = Machine.objects.get_current()
        self.archiveresult.iface = NetworkInterface.objects.get_current()
        self.archiveresult.save()
        self.archiveresult.write_indexes()
    
    def extract(self, url: str, binaries: dict, daemons: dict, cwd: Path, config: dict):
        proc = subprocess.run(self.archiveresult.cmd, cwd=self.archiveresult.cwd, env=os.environ.update(binaries), timeout=self.timeout, shell=True, capture_output=True, text=True)
        self.archiveresult.stdout = proc.stdout
        self.archiveresult.stderr = proc.stderr
        self.archiveresult.returncode = proc.returncode
        self.archiveresult.save()
        self.archiveresult.write_indexes()
        
    def determine_status(self):
        if self.archiveresult.returncode == 29:
            return self.archiveresult.StatusChoices.BACKOFF, timezone.now() + timedelta(seconds=self.timeout)
        elif self.archiveresult.returncode == 50:
            return self.archiveresult.StatusChoices.SKIPPED, None
        else:
            return self.archiveresult.StatusChoices.FAILED, None

    def collect_outputs(self, cwd: Path):
        for file in cwd.rglob('*'):
            path = file.relative_to(cwd)
            os.chmod(file, 0o644)
            #os.chown(file, ARCHIVEBOX_UID, ARCHIVEBOX_GID)
            
            self.archiveresult.outputs.append({
                'type': 'FILE',
                'path': file.relative_to(cwd),
                'size': file.stat().st_size,
                'ext': file.suffix,
                'mimetype': mimetypes.guess_type(file)[0],
                'sha256': hashlib.sha256(file.read_bytes()).hexdigest(),
                'blake3': hashlib.blake3(file.read_bytes()).hexdigest(),
                'created_at': file.stat().st_ctime,
                'modified_at': file.stat().st_mtime,
                'symlinks': [
                    'screenshot.png',
                    'example.com',
                ]
            })
            outlinks = parse_outlinks(file)
            if outlinks:
                self.archiveresult.outputs.append({
                    'type': 'OUTLINK',
                    'url': outlink.target,
                    'selector': outlink.selector,
                    'text': outlink.text,
                })

            if path.endswith('favicon.ico'):
                self.archiveresult.outputs.append({
                    'type': 'FAVICON',
                    'symlinks': {
                        'favicon': output_file['path'],
                        'favicon.ico': output_file['path'],
                        'favicon.png': output_file['path'].with_suffix('.png'),
                    },
                    'path': output_file['path'],
                })
            if path.endswith('.pdf'):
                self.archiveresult.outputs.append({
                    'type': 'PDF',
                    'path': file.relative_to(cwd),
                })
                
            if 'text/plain' in mimetypes.guess_type(file):
                self.archiveresult.outputs.append({
                    'type': 'SEARCHTEXT',
                    'path': file.relative_to(self.archiveresult.OUTPUT_DIR),
                    'archiveresult_id': self.archiveresult.id,
                })
    
    def after_extract(self, error: Exception | None=None):
        status, retry_at = self.determine_status()

        
        self.archiveresult.error = f'{type(error).__name__}: {error}' if error else None
        self.archiveresult.status = self.archiveresult.StatusChoices.FAILED if error else self.archiveresult.StatusChoices.SUCCEEDED
        self.archiveresult.retry_at = None
        self.archiveresult.end_ts = timezone.now()
        self.archiveresult.output = self.archiveresult.outputs[0].path
        self.archiveresult.save()
        self.archiveresult.write_indexes()