Merge pull request #261 from bellingcat/wacz_separate_profile

Wacz minor adjustments

docker-compose.yaml

@@ -1,4 +1,3 @@
-version: '3.8'
 
 services:
   auto-archiver:
@@ -10,7 +9,4 @@ services:
     volumes:
       - ./secrets:/app/secrets
       - ./local_archive:/app/local_archive
-    environment:
-      - WACZ_ENABLE_DOCKER=true
-      - RUNNING_IN_DOCKER=true
     command: --config secrets/orchestration.yaml

src/auto_archiver/core/module.py

@@ -5,6 +5,7 @@ by handling user configuration, validating the steps properties, and implementin
 """
 
 from __future__ import annotations
+import subprocess
 
 from dataclasses import dataclass
 from typing import List, TYPE_CHECKING, Type
@@ -17,7 +18,7 @@ import os
 from os.path import join
 from loguru import logger
 import auto_archiver
-from auto_archiver.core.consts import DEFAULT_MANIFEST, MANIFEST_FILE
+from auto_archiver.core.consts import DEFAULT_MANIFEST, MANIFEST_FILE, SetupError
 
 if TYPE_CHECKING:
     from .base_module import BaseModule
@@ -220,9 +221,9 @@ class LazyBaseModule:
                 if not check(dep):
                     logger.error(
                         f"Module '{self.name}' requires external dependency '{dep}' which is not available/setup. \
-                                 Have you installed the required dependencies for the '{self.name}' module? See the README for more information."
+                                 Have you installed the required dependencies for the '{self.name}' module? See the documentation for more information."
                     )
-                    exit(1)
+                    raise SetupError()
 
         def check_python_dep(dep):
             # first check if it's a module:
@@ -241,8 +242,22 @@ class LazyBaseModule:
 
             return find_spec(dep)
 
+        def check_bin_dep(dep):
+            dep_exists = shutil.which(dep)
+
+            if dep == "docker":
+                if os.environ.get("RUNNING_IN_DOCKER"):
+                    # this is only for the WACZ enricher, which requires docker
+                    # if we're already running in docker then we don't need docker
+                    return True
+
+                # check if docker daemon is running
+                return dep_exists and subprocess.run(["docker", "ps", "-q"]).returncode == 0
+
+            return dep_exists
+
         check_deps(self.dependencies.get("python", []), check_python_dep)
-        check_deps(self.dependencies.get("bin", []), lambda dep: shutil.which(dep))
+        check_deps(self.dependencies.get("bin", []), check_bin_dep)
 
         logger.debug(f"Loading module '{self.display_name}'...")
 

src/auto_archiver/modules/wacz_extractor_enricher/__manifest__.py

@@ -11,7 +11,7 @@
     "configs": {
         "profile": {
             "default": None,
-            "help": "browsertrix-profile (for profile generation see https://github.com/webrecorder/browsertrix-crawler#creating-and-using-browser-profiles).",
+            "help": "browsertrix-profile (for profile generation see https://crawler.docs.browsertrix.com/user-guide/browser-profiles/).",
         },
         "docker_commands": {"default": None, "help": "if a custom docker invocation is needed"},
         "timeout": {"default": 120, "help": "timeout for WACZ generation in seconds", "type": "int"},
@@ -40,14 +40,27 @@
     Creates .WACZ archives of web pages using the `browsertrix-crawler` tool, with options for media extraction and screenshot saving.
     [Browsertrix-crawler](https://crawler.docs.browsertrix.com/user-guide/) is a headless browser-based crawler that archives web pages in WACZ format.
 
-    ### Features
+    ## Setup
+
+    **Docker**
+    If you are using the Docker file to run Auto Archiver (recommended), then everything is set up and you can use WACZ out of the box!
+    Otherwise, if you are using a local install of Auto Archiver (e.g. pip or dev install), then you will need to install Docker and run 
+    the docker daemon to be able to run the `browsertrix-crawler` tool.
+
+    **Browsertrix Profiles**
+    A browsertrix profile is a custom browser profile (login information, browser extensions, etc.) that can be used to archive private or dynamic content.
+    You can run the WACZ Enricher without a profile, but for more resilient archiving, it is recommended to create a profile. See the [Browsertrix documentation](https://crawler.docs.browsertrix.com/user-guide/browser-profiles/)
+    for more information.
+
+    ** Docker in Docker **
+    If you are running Auto Archiver within a Docker container, you will need to enable Docker in Docker to run the `browsertrix-crawler` tool.
+    This can be done by setting the `WACZ_ENABLE_DOCKER` environment variable to `1`.
+
+    ## Features
     - Archives web pages into .WACZ format using Docker or direct invocation of `browsertrix-crawler`.
     - Supports custom profiles for archiving private or dynamic content.
     - Extracts media (images, videos, audio) and screenshots from the archive, optionally adding them to the enrichment pipeline.
     - Generates metadata from the archived page's content and structure (e.g., titles, text).
 
-    ### Notes
-    - Requires Docker for running `browsertrix-crawler` .
-    - Configurable via parameters for timeout, media extraction, screenshots, and proxy settings.
     """,
 }

src/auto_archiver/modules/wacz_extractor_enricher/wacz_extractor_enricher.py

@@ -24,7 +24,8 @@ class WaczExtractorEnricher(Enricher, Extractor):
         self.use_docker = os.environ.get("WACZ_ENABLE_DOCKER") or not os.environ.get("RUNNING_IN_DOCKER")
         self.docker_in_docker = os.environ.get("WACZ_ENABLE_DOCKER") and os.environ.get("RUNNING_IN_DOCKER")
 
-        self.cwd_dind = f"/crawls/crawls{random_str(8)}"
+        self.crawl_id = random_str(8)
+        self.cwd_dind = f"/crawls/crawls{self.crawl_id}"
         self.browsertrix_home_host = os.environ.get("BROWSERTRIX_HOME_HOST")
         self.browsertrix_home_container = os.environ.get("BROWSERTRIX_HOME_CONTAINER") or self.browsertrix_home_host
         # create crawls folder if not exists, so it can be safely removed in cleanup
@@ -50,7 +51,7 @@ class WaczExtractorEnricher(Enricher, Extractor):
 
         url = to_enrich.get_url()
 
-        collection = random_str(8)
+        collection = self.crawl_id
         browsertrix_home_host = self.browsertrix_home_host or os.path.abspath(self.tmp_dir)
         browsertrix_home_container = self.browsertrix_home_container or browsertrix_home_host
 
@@ -102,10 +103,11 @@ class WaczExtractorEnricher(Enricher, Extractor):
                 ] + cmd
 
             if self.profile:
-                profile_fn = os.path.join(browsertrix_home_container, "profile.tar.gz")
+                profile_file = f"profile-{self.crawl_id}.tar.gz"
+                profile_fn = os.path.join(browsertrix_home_container, profile_file)
                 logger.debug(f"copying {self.profile} to {profile_fn}")
                 shutil.copyfile(self.profile, profile_fn)
-                cmd.extend(["--profile", os.path.join("/crawls", "profile.tar.gz")])
+                cmd.extend(["--profile", os.path.join("/crawls", profile_file)])
 
         else:
             logger.debug(f"generating WACZ without Docker for {url=}")

tests/enrichers/test_wacz_enricher.py

@@ -4,6 +4,7 @@ from zipfile import ZipFile
 import pytest
 
 from auto_archiver.core import Metadata, Media
+from auto_archiver.core.consts import SetupError
 
 
 @pytest.fixture
@@ -22,6 +23,15 @@ def wacz_enricher(setup_module, mock_binary_dependencies):
     return wacz
 
 
+def test_raises_error_without_docker_installed(setup_module, mocker, caplog):
+    # pretend that docker isn't installed
+    mocker.patch("shutil.which").return_value = None
+    with pytest.raises(SetupError):
+        setup_module("wacz_extractor_enricher", {})
+
+    assert "requires external dependency 'docker' which is not available/setup" in caplog.text
+
+
 def test_setup_without_docker(wacz_enricher, mocker):
     mocker.patch.dict(os.environ, {"RUNNING_IN_DOCKER": "1"}, clear=True)
     wacz_enricher.setup()

tests/test_modules.py

@@ -1,6 +1,7 @@
 import pytest
 from auto_archiver.core.module import ModuleFactory, LazyBaseModule
 from auto_archiver.core.base_module import BaseModule
+from auto_archiver.core.consts import SetupError
 
 
 @pytest.fixture
@@ -25,11 +26,9 @@ def test_python_dependency_check(example_module):
     # monkey patch the manifest to include a nonexistnet dependency
     example_module.manifest["dependencies"]["python"] = ["does_not_exist"]
 
-    with pytest.raises(SystemExit) as load_error:
+    with pytest.raises(SetupError):
         example_module.load({})
 
-    assert load_error.value.code == 1
-
 
 def test_binary_dependency_check(example_module):
     # example_module requires ffmpeg, which is not installed