Data Aggregator Rewrite (openwpm#753)

* First steps in the rewrite * Fixed import paths * One giant refactor * Fixing tests * Adding mypy * Removed mypy from pre-commit workflow * First draft on DataAggregator * Wrote a DataAggregator that starts and shuts down * Created tests and added more empty types * Got demo.py working * Created sql_provider * Cleaned up imports in TaskManager * Added async * Fixed minor bugs * First steps at porting arrow * Introduced TableName and different Task handling * Added more failing tests * First first completes others don't * It works * Started working on arrow_provider * Implemented ArrowProvider * Added logger fixture * Fixed test_storage_controller * Fixing OpenWPMTest.visit() * Moved test/storage_providers to test/storage * Fixing up tests * Moved automation to openwpm * Readded datadir to .gitignore * Ran repin.sh * Fixed formatting * Let's see if this works * Fixed imports * Got arrow_memory_provider working * Starting to rewrite tests * Setting up fixtures * Attempting to fix all the tests * Still fixing tests * Broken content saving * Added node * Fixed screenshot tests * Fixing more tests * Fixed tests * Implemented local_storage.py * Cleaned up flush_cache * Fixing more tests * Wrote test for LocalArrowProvider * Introduced tests for local_storage_provider.py * Asserting test dir is empty * Creating subfolder for different aggregators * New depencies and init() * Everything is terribly broken * Figured out finalize_visit_id * Running two event loops kinda works??? * Rearming the event * Introduced mypy * Downgraded black in pre-commit * Modifying the database directly * Fixed formatting * Made mypy a lil stricter * Fixing docs and config printing * Realising I've been using the wrong with * Trying to figure arrow_storage * Moving lock initialization in in_memory_storage * Fixing tests * Fixing up tests and adding more typechecking * Fixed num_browsers in test_cache_hits_recorded * Parametrized unstructured * String fix * Added failing test * New test * Review changes with Steven * Fixed repin.sh and test_arrow_cache * Minor change * Fixed prune-environment.py * Removing references to DataAggregator * Fixed test_seed_persistance * More paths * Fixed test display shutdown * Made cache test more robust * Update crawler.py Co-authored-by: Steven Englehardt <senglehardt@mozilla.com> * Slimming down ManagerParams * Fixing more tests * Update test/storage/test_storage_controller.py Co-authored-by: Steven Englehardt <senglehardt@mozilla.com> * Purging references to DataAggregator * Reverted changes to .travis.yml * Demo.py saves locally again * Readjusting test paths * Expanded comment on initialize to reference openwpm#846 * Made token optional in finalize_visit_id * Simplified test paramtetrization * Fixed callback semantics change * Removed test_parse_http_stack_trace_str * Added DataSocket * WIP need to fix path encoding * Fixed path encoding * Added task and crawl to schema * Fixed paths in GitHub actions * Refactored completion handling * Fix tests * Trying to fix tests on CI * Removed redundant setting of tag * Removing references to S3 * Purging more DataAggregator references * Craking up logging to figure out test failure * Moved test_values into a fixture * Fixing GcpUnstructuredProvider * Fixed paths for future crawls * Renamed sqllite to official sqlite * Restored demo.py * Update openwpm/commands/profile_commands.py Co-authored-by: Georgia Kokkinou <geor5ko@gmail.com> * Restored previous behaviour of DumpProfileCommand Co-authored-by: Georgia Kokkinou <geor5ko@gmail.com> * Removed leftovers * Cleaned up comments * Expanded lock check * Fixed more stuff * More comment updates * Update openwpm/socket_interface.py Co-authored-by: Georgia Kokkinou <geor5ko@gmail.com> * Removed outdated comment * Using config_encoder * Renamed tar_location to tar_path * Removed references to database_name in docs * Cleanup * Moved screenshot_path and source_dump_path to ManagerParamsInternal * Fixed imports * Fixing up comments * Fixing up comments * More docs * updated dependencies * Fixed test_task_manager * Reupgraded to python 3.9.1 * Restoring crawl_reference in mp_logger * Removed unused imports * Apply suggestions from code review Co-authored-by: Steven Englehardt <senglehardt@mozilla.com> * Cleaned up socket handling * Fixed TaskManager.__exit__ * Moved validation code into config.py * Removed comment * Removed comment * Removed comment Co-authored-by: Steven Englehardt <senglehardt@mozilla.com> Co-authored-by: Georgia Kokkinou <geor5ko@gmail.com>
Zaxeli · Feb 22, 2021 · 507d769 · 507d769
1 parent 7efd249
commit 507d769
Show file tree

Hide file tree

Showing 96 changed files with 4,422 additions and 3,095 deletions.
diff --git a/.github/workflows/run-tests.yaml b/.github/workflows/run-tests.yaml
@@ -48,7 +48,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        test-groups: ["test_[a-e]*", "test_[f-h]*", "test_[i-r,t-z]*", "test_[s]*"]
+        test-groups: ["test/test_[a-e]*", "test/test_[f-h]*", "test/test_[i-r,t-z]*", "test/test_[s]*", "test/storage/*"]
       fail-fast: false
     steps:
     # All of these steps are just setup, maybe we should wrap them in an action

diff --git a/.gitignore b/.gitignore
@@ -94,3 +94,5 @@ openwpm/Extension/firefox/dist
 openwpm/Extension/firefox/openwpm.xpi
 openwpm/Extension/firefox/src/content.js
 openwpm/Extension/firefox/src/feature.js
+
+datadir
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -4,7 +4,16 @@ repos:
     hooks:
     -   id: isort
   - repo: https://github.com/psf/black
-    rev: 20.8b1 # Replace by any tag/version: https://github.com/psf/black/tags
+    rev: 20.8b1
     hooks:
       - id: black
-        language_version: python3 # Should be a command that runs python3.6+
+        language_version: python3
+  - repo: https://github.com/pre-commit/mirrors-mypy
+    rev: v0.790
+    hooks:
+      - id: mypy
+        additional_dependencies: [pytest]
+        # We may need to add more and more dependencies here, as pre-commit
+        # runs in an environment without our dependencies
+
+
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -85,7 +85,6 @@ OpenWPM's tests are build on [pytest](https://docs.pytest.org/en/latest/). Execu
 in the test directory to run all tests:
 
     $ conda activate openwpm
-    $ cd test
     $ py.test -vv
 
 See the [pytest docs](https://docs.pytest.org/en/latest/) for more information on selecting

diff --git a/README.md b/README.md
@@ -12,25 +12,25 @@ the instrumentation section below for more details.
 
 Table of Contents <!-- omit in toc -->
 ------------------
-* [Installation](#installation)
-  * [Pre-requisites](#pre-requisites)
-  * [Install](#install)
-  * [Mac OSX](#mac-osx)
-* [Quick Start](#quick-start)
-* [Troubleshooting](#troubleshooting)
-* [Advice for Measurement Researchers](#advice-for-measurement-researchers)
-* [Developer instructions](#developer-instructions)
-* [Instrumentation and Configuration](#instrumentation-and-configuration)
-* [Persistence Types](#persistence-types)
-    * [Local Databases](#local-databases)
-    * [Parquet on Amazon S3](#parquet-on-amazon-s3)
-* [Docker Deployment for OpenWPM](#docker-deployment-for-openwpm)
-  * [Building the Docker Container](#building-the-docker-container)
-  * [Running Measurements from inside the Container](#running-measurements-from-inside-the-container)
-  * [MacOS GUI applications in Docker](#macos-gui-applications-in-docker)
-* [Citation](#citation)
-* [License](#license)
 
+- [Installation](#installation)
+  - [Pre-requisites](#pre-requisites)
+  - [Install](#install)
+  - [Mac OSX](#mac-osx)
+- [Quick Start](#quick-start)
+- [Troubleshooting](#troubleshooting)
+- [Advice for Measurement Researchers](#advice-for-measurement-researchers)
+- [Developer instructions](#developer-instructions)
+- [Instrumentation and Configuration](#instrumentation-and-configuration)
+- [Storage](#storage)
+  - [Local Storage](#local-storage)
+  - [Remote storage](#remote-storage)
+- [Docker Deployment for OpenWPM](#docker-deployment-for-openwpm)
+  - [Building the Docker Container](#building-the-docker-container)
+  - [Running Measurements from inside the Container](#running-measurements-from-inside-the-container)
+  - [MacOS GUI applications in Docker](#macos-gui-applications-in-docker)
+- [Citation](#citation)
+- [License](#license)
 
 Installation
 ------------
@@ -178,40 +178,52 @@ If you want to contribute to OpenWPM have a look at our [CONTRIBUTING.md](./CONT
 
 Instrumentation and Configuration
 -------------------------------
+
 OpenWPM provides a breadth of configuration options which can be found
 in [Configuration.md](docs/Configuration.md)
 More detail on the output is available [below](#persistence-types).
 
-Persistence Types
+Storage
 ------------
 
-#### Local Databases
-By default OpenWPM saves all data locally on disk in a variety of formats.
-Most of the instrumentation saves to a SQLite database specified
-by `manager_params.database_name` in the main output directory. Response
-bodies are saved in a LevelDB database named `content.ldb`, and are keyed by
-the hash of the content. In addition, the browser commands that dump page
-source and save screenshots save them in the `sources` and `screenshots`
-subdirectories of the main output directory. The SQLite schema
-specified by: `openwpm/DataAggregator/schema.sql`. You can specify additional tables
-inline by sending a `create_table` message to the data aggregator.
-
-#### Parquet on Amazon S3
-As an option, OpenWPM can save data directly to an Amazon S3 bucket as a
-Parquet Dataset. This is currently experimental and hasn't been thoroughly
-tested. Screenshots, and page source saving is not currently supported and
-will still be stored in local databases and directories. To enable S3
-saving specify the following configuration parameters in `manager_params`:
-* Persistence Type: `manager_params.output_format = 's3'`
-* S3 bucket name: `manager_params.s3_bucket = 'openwpm-test-crawl'`
-* Directory within S3 bucket: `manager_params.s3_directory = '2018-09-09_test-crawl-new'`
-
-In order to save to S3 you must have valid access credentials stored in
-`~/.aws`. We do not currently allow you to specify an alternate storage
-location.
-
-**NOTE:** The schemas should be kept in sync with the exception of
-output-specific columns (e.g., `instance_id` in the S3 output). You can compare
+OpenWPM distinguishes between two types of data, structured and unstructured.
+Structured data is all data captured by the instrumentation or emitted by the platform.
+Generally speaking all data you download is unstructured data.
+
+For each of the data classes we offer a variety of storage providers, and you are encouraged
+to implement your own, should the provided backends not be enough for you.
+
+We have an outstanding issue to enable saving content generated by commands, such as
+screenshots and page dumps to unstructured storage (see [#232](https://github.com/mozilla/OpenWPM/issues/232)).  
+For now, they get saved to `manager_params.data_directory`.
+
+### Local Storage
+
+For storing structured data locally we offer two StorageProviders:
+
+- The SQLiteStorageProvider which writes all data into a SQLite database
+  - This is the recommended approach for getting started as the data is easily explorable
+- The LocalArrowProvider which stores the data into Parquet files.
+  - This method integrates well with NumPy/Pandas
+  - It might be harder to ad-hoc process
+
+For storing unstructured data locally we also offer two solutions:
+
+- The LevelDBProvider which stores all data into a LevelDB
+  - This is the recommended approach
+- The LocalGzipProvider that gzips and stores the files individually on disk
+  - Please note that file systems usually don't like thousands of files in one folder
+  - Use with care or for single site visits
+
+### Remote storage
+
+When running in the cloud, saving records to disk is not a reasonable thing to do.
+So we offer a remote StorageProviders for S3 (See [#823](https://github.com/mozilla/OpenWPM/issues/823)) and GCP.
+Currently, all remote StorageProviders write to the respective object storage service (S3/GCS).
+The structured providers use the Parquet format.
+
+**NOTE:** The Parquet and SQL schemas should be kept in sync except
+output-specific columns (e.g., `instance_id` in the Parquet output). You can compare
 the two schemas by running
 `diff -y openwpm/DataAggregator/schema.sql openwpm/DataAggregator/parquet_schema.py`.
 
@@ -238,7 +250,7 @@ Docker service.
 __Step 2:__ to build the image, run the following command from a terminal
 within the root OpenWPM directory:
 
-```
+```bash
     docker build -f Dockerfile -t openwpm .
 ```
 
@@ -253,7 +265,7 @@ X-server. You can do this by running: `xhost +local:docker`
 
 Then you can run the demo script using:
 
-```
+```bash
     mkdir -p docker-volume && docker run -v $PWD/docker-volume:/opt/Desktop \
     -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix --shm-size=2g \
     -it openwpm python3 /opt/OpenWPM/demo.py

diff --git a/crawler.py b/crawler.py
@@ -4,42 +4,53 @@
 import signal
 import sys
 import time
+from pathlib import Path
 from threading import Lock
 from typing import Any, Callable, List
 
-import boto3
 import sentry_sdk
 
+from openwpm import mp_logger
 from openwpm.command_sequence import CommandSequence
 from openwpm.config import BrowserParams, ManagerParams
-from openwpm.mp_logger import parse_config_from_env
+from openwpm.storage.cloud_storage.gcp_storage import (
+    GcsStructuredProvider,
+    GcsUnstructuredProvider,
+)
 from openwpm.task_manager import TaskManager
 from openwpm.utilities import rediswq
-from test.utilities import LocalS3Session, local_s3_bucket
 
 # Configuration via environment variables
+# Crawler specific config
 REDIS_HOST = os.getenv("REDIS_HOST", "redis-box")
 REDIS_QUEUE_NAME = os.getenv("REDIS_QUEUE_NAME", "crawl-queue")
+MAX_JOB_RETRIES = int(os.getenv("MAX_JOB_RETRIES", "2"))
+DWELL_TIME = int(os.getenv("DWELL_TIME", "10"))
+TIMEOUT = int(os.getenv("TIMEOUT", "60"))
+
+# Storage Provider Params
 CRAWL_DIRECTORY = os.getenv("CRAWL_DIRECTORY", "crawl-data")
-S3_BUCKET = os.getenv("S3_BUCKET", "openwpm-crawls")
+GCS_BUCKET = os.getenv("GCS_BUCKET", "openwpm-crawls")
+GCP_PROJECT = os.getenv("GCP_PROJECT", "")
+AUTH_TOKEN = os.getenv("GCP_AUTH_TOKEN", "cloud")
+
+# Browser Params
 DISPLAY_MODE = os.getenv("DISPLAY_MODE", "headless")
 HTTP_INSTRUMENT = os.getenv("HTTP_INSTRUMENT", "1") == "1"
 COOKIE_INSTRUMENT = os.getenv("COOKIE_INSTRUMENT", "1") == "1"
 NAVIGATION_INSTRUMENT = os.getenv("NAVIGATION_INSTRUMENT", "1") == "1"
 JS_INSTRUMENT = os.getenv("JS_INSTRUMENT", "1") == "1"
 CALLSTACK_INSTRUMENT = os.getenv("CALLSTACK_INSTRUMENT", "1") == "1"
-JS_INSTRUMENT_SETTINGS = os.getenv(
-    "JS_INSTRUMENT_SETTINGS", '["collection_fingerprinting"]'
+JS_INSTRUMENT_SETTINGS = json.loads(
+    os.getenv("JS_INSTRUMENT_SETTINGS", '["collection_fingerprinting"]')
 )
+
 SAVE_CONTENT = os.getenv("SAVE_CONTENT", "")
 PREFS = os.getenv("PREFS", None)
-DWELL_TIME = int(os.getenv("DWELL_TIME", "10"))
-TIMEOUT = int(os.getenv("TIMEOUT", "60"))
-SENTRY_DSN = os.getenv("SENTRY_DSN", None)
-LOGGER_SETTINGS = parse_config_from_env()
-MAX_JOB_RETRIES = int(os.getenv("MAX_JOB_RETRIES", "2"))
 
-JS_INSTRUMENT_SETTINGS = json.loads(JS_INSTRUMENT_SETTINGS)
+
+SENTRY_DSN = os.getenv("SENTRY_DSN", None)
+LOGGER_SETTINGS = mp_logger.parse_config_from_env()
 
 if CALLSTACK_INSTRUMENT is True:
     # Must have JS_INSTRUMENT True for CALLSTACK_INSTRUMENT to work
@@ -74,29 +85,38 @@
         browser_params[i].prefs = json.loads(PREFS)
 
 # Manager configuration
-manager_params.data_directory = "~/Desktop/%s/" % CRAWL_DIRECTORY
-manager_params.log_directory = "~/Desktop/%s/" % CRAWL_DIRECTORY
-manager_params.output_format = "s3"
-manager_params.s3_bucket = S3_BUCKET
-manager_params.s3_directory = CRAWL_DIRECTORY
-
-# Allow the use of localstack's mock s3 service
-S3_ENDPOINT = os.getenv("S3_ENDPOINT")
-if S3_ENDPOINT:
-    boto3.DEFAULT_SESSION = LocalS3Session(endpoint_url=S3_ENDPOINT)
-    manager_params.s3_bucket = local_s3_bucket(boto3.resource("s3"), name=S3_BUCKET)
-
+manager_params.data_directory = Path("~/Desktop/") / CRAWL_DIRECTORY
+manager_params.log_directory = Path("~/Desktop/") / CRAWL_DIRECTORY
+
+structured = GcsStructuredProvider(
+    project=GCP_PROJECT,
+    bucket_name=GCS_BUCKET,
+    base_path=CRAWL_DIRECTORY,
+    token=AUTH_TOKEN,
+)
+unstructured = GcsUnstructuredProvider(
+    project=GCP_PROJECT,
+    bucket_name=GCS_BUCKET,
+    base_path=CRAWL_DIRECTORY + "/data",
+    token=AUTH_TOKEN,
+)
 # Instantiates the measurement platform
 # Commands time out by default after 60 seconds
-manager = TaskManager(manager_params, browser_params, logger_kwargs=LOGGER_SETTINGS)
+manager = TaskManager(
+    manager_params,
+    browser_params,
+    structured,
+    unstructured,
+    logger_kwargs=LOGGER_SETTINGS,
+)
 
 # At this point, Sentry should be initiated
 if SENTRY_DSN:
     # Add crawler.py-specific context
     with sentry_sdk.configure_scope() as scope:
         # tags generate breakdown charts and search filters
         scope.set_tag("CRAWL_DIRECTORY", CRAWL_DIRECTORY)
-        scope.set_tag("S3_BUCKET", S3_BUCKET)
+        scope.set_tag("GCS_BUCKET", GCS_BUCKET)
         scope.set_tag("DISPLAY_MODE", DISPLAY_MODE)
         scope.set_tag("HTTP_INSTRUMENT", HTTP_INSTRUMENT)
         scope.set_tag("COOKIE_INSTRUMENT", COOKIE_INSTRUMENT)
@@ -108,9 +128,10 @@
         scope.set_tag("DWELL_TIME", DWELL_TIME)
         scope.set_tag("TIMEOUT", TIMEOUT)
         scope.set_tag("MAX_JOB_RETRIES", MAX_JOB_RETRIES)
-        scope.set_tag("CRAWL_REFERENCE", "%s/%s" % (S3_BUCKET, CRAWL_DIRECTORY))
+        scope.set_tag("CRAWL_REFERENCE", "%s/%s" % (GCS_BUCKET, CRAWL_DIRECTORY))
         # context adds addition information that may be of interest
-        scope.set_context("PREFS", PREFS)
+        if PREFS:
+            scope.set_context("PREFS", json.loads(PREFS))
         scope.set_context(
             "crawl_config",
             {
@@ -159,9 +180,9 @@ def get_job_completion_callback(
     job_queue: rediswq.RedisWQ,
     job: bytes,
 ) -> Callable[[bool], None]:
-    def callback(sucess: bool) -> None:
+    def callback(success: bool) -> None:
         with unsaved_jobs_lock:
-            if sucess:
+            if success:
                 logger.info("Job %r is done", job)
                 job_queue.complete(job)
             else: