Metadata-Version: 2.4
Name: silkworm-rs
Version: 0.4.8
Summary: Async-first web scraping framework built on wreq and scraper-rs
Project-URL: Homepage, https://github.com/BitingSnakes/silkworm
Project-URL: Repository, https://github.com/BitingSnakes/silkworm
Project-URL: Issues, https://github.com/BitingSnakes/silkworm/issues
Author-email: Yehor Smoliakov <egorsmkv@gmail.com>
License: MIT License
        
        Copyright (c) 2025 Yehor Smoliakov
        
        Permission is hereby granted, free of charge, to any person obtaining a copy
        of this software and associated documentation files (the "Software"), to deal
        in the Software without restriction, including without limitation the rights
        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
        copies of the Software, and to permit persons to whom the Software is
        furnished to do so, subject to the following conditions:
        
        The above copyright notice and this permission notice shall be included in all
        copies or substantial portions of the Software.
        
        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
        SOFTWARE.
License-File: LICENSE
Keywords: asyncio,crawler,html,http,spider,web-scraping
Classifier: Development Status :: 5 - Production/Stable
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3 :: Only
Classifier: Programming Language :: Python :: 3.13
Classifier: Programming Language :: Python :: 3.14
Classifier: Topic :: Internet :: WWW/HTTP
Classifier: Topic :: Internet :: WWW/HTTP :: Indexing/Search
Classifier: Topic :: Software Development :: Libraries
Classifier: Typing :: Typed
Requires-Python: <3.15,>=3.13
Requires-Dist: charset-normalizer==3.4.4
Requires-Dist: logly==0.1.6
Requires-Dist: pyopenssl>=26.0.0
Requires-Dist: requests>=2.33.0
Requires-Dist: rxml==2.4.1
Requires-Dist: scraper-rust>=0.4.3
Requires-Dist: wreq>=0.10.0
Provides-Extra: avro
Requires-Dist: fastavro==1.12.1; extra == 'avro'
Provides-Extra: cassandra
Requires-Dist: cassandra-driver==3.29.3; (sys_platform != 'win32' and python_version < '3.14') and extra == 'cassandra'
Provides-Extra: cdp
Requires-Dist: websockets==16.0; extra == 'cdp'
Provides-Extra: couchdb
Requires-Dist: aiocouch==4.0.1; extra == 'couchdb'
Provides-Extra: duckdb
Requires-Dist: duckdb==1.5.1; extra == 'duckdb'
Provides-Extra: dynamodb
Requires-Dist: aioboto3==15.5.0; extra == 'dynamodb'
Provides-Extra: elasticsearch
Requires-Dist: elasticsearch==9.3.0; extra == 'elasticsearch'
Provides-Extra: excel
Requires-Dist: openpyxl==3.1.5; extra == 'excel'
Provides-Extra: ftp
Requires-Dist: aioftp==0.27.2; extra == 'ftp'
Provides-Extra: gsheets
Requires-Dist: google-api-python-client==2.193.0; extra == 'gsheets'
Requires-Dist: google-auth==2.49.1; extra == 'gsheets'
Provides-Extra: memray
Requires-Dist: memray==1.19.2; extra == 'memray'
Provides-Extra: mongodb
Requires-Dist: motor==3.7.1; extra == 'mongodb'
Provides-Extra: msgpack
Requires-Dist: ormsgpack==1.12.2; extra == 'msgpack'
Provides-Extra: mysql
Requires-Dist: aiomysql==0.3.2; extra == 'mysql'
Provides-Extra: onionlink
Requires-Dist: onionlink==0.1.2; extra == 'onionlink'
Provides-Extra: polars
Requires-Dist: polars==1.39.3; extra == 'polars'
Provides-Extra: postgresql
Requires-Dist: asyncpg==0.31.0; extra == 'postgresql'
Provides-Extra: rsloop
Requires-Dist: rsloop==0.1.14; extra == 'rsloop'
Provides-Extra: s3
Requires-Dist: opendal==0.46.0; extra == 's3'
Provides-Extra: sftp
Requires-Dist: asyncssh==2.22.0; extra == 'sftp'
Provides-Extra: snowflake
Requires-Dist: snowflake-connector-python==4.4.0; extra == 'snowflake'
Provides-Extra: taskiq
Requires-Dist: taskiq==0.12.1; extra == 'taskiq'
Provides-Extra: trio
Requires-Dist: trio-asyncio==0.15.0; extra == 'trio'
Requires-Dist: trio==0.33.0; extra == 'trio'
Provides-Extra: uvloop
Requires-Dist: uvloop==0.22.1; extra == 'uvloop'
Provides-Extra: vortex
Requires-Dist: vortex-data==0.67.0; extra == 'vortex'
Provides-Extra: winloop
Requires-Dist: winloop==0.6.0; extra == 'winloop'
Provides-Extra: yaml
Requires-Dist: pyyaml==6.0.3; extra == 'yaml'
Description-Content-Type: text/markdown

# silkworm-rs

[![PyPI - Version](https://img.shields.io/pypi/v/silkworm-rs)](https://pypi.org/project/silkworm-rs/)
[![Tests](https://github.com/BitingSnakes/silkworm/actions/workflows/tests.yml/badge.svg)](https://github.com/BitingSnakes/silkworm/actions/workflows/tests.yml)
[![PyPI Downloads](https://static.pepy.tech/personalized-badge/silkworm-rs?period=total&units=INTERNATIONAL_SYSTEM&left_color=BLACK&right_color=GREEN&left_text=downloads)](https://pepy.tech/projects/silkworm-rs)

Async-first web scraping framework built on [wreq](https://github.com/0x676e67/wreq-python) (HTTP with browser impersonation) and [scraper-rs](https://github.com/RustedBytes/scraper-rs) (fast HTML parsing). Silkworm gives you a minimal Spider/Request/Response model, middlewares, and pipelines so you can script quick scrapes or build larger crawlers without boilerplate.

> **NEW**: Use [silkworm-mcp](https://github.com/BitingSnakes/silkworm-mcp) to build scrapers.

## Features
- Async engine with configurable concurrency, bounded queue backpressure (defaults to `concurrency * 10`), and per-request timeouts.
- wreq-powered HTTP client: browser impersonation, redirect following with loop detection, query merging, and proxy support via `request.meta["proxy"]`.
- Optional OnionLink client integration for scraping Tor v3 `.onion` sites without routing through wreq.
- Typed spiders and callbacks that can return items or `Request` objects; `HTMLResponse` ships helper methods plus `Response.follow` to reuse callbacks.
- Middlewares: User-Agent rotation/default, proxy rotation, retry with exponential backoff + optional sleep codes, flexible delays (fixed/random/custom), `SkipNonHTMLMiddleware` to drop non-HTML callbacks, and `CloudflareCrawlMiddleware` for Browser Rendering crawl jobs.
- Pipelines: JSON Lines, SQLite, XML (nested data preserved), and CSV (flattens dicts and lists) out of the box.
- Structured logging via `logly` (`SILKWORM_LOG_LEVEL=DEBUG`), plus periodic/final crawl statistics (requests/sec, queue size, memory, seen URLs).

## Installation

From PyPI with pip:

```bash
pip install silkworm-rs
```

From PyPI with uv (recommended for faster installs):

```bash
uv pip install silkworm-rs
# or if using uv's project management:
uv add silkworm-rs
```

From source:

```bash
uv venv  # install uv from https://docs.astral.sh/uv/getting-started/ if needed
source .venv/bin/activate  # Windows: .venv\Scripts\activate
uv pip install -e .
```

Targets Python 3.13+; dependencies are pinned in `pyproject.toml`.

## Quick start
Define a spider by subclassing `Spider`, implementing `parse`, and yielding items or follow-up `Request` objects. This example writes quotes to `data/quotes.jl` and enables basic user agent, retry, and non-HTML filtering middlewares.

```python
from silkworm import HTMLResponse, Response, Spider, run_spider
from silkworm.middlewares import (
    RetryMiddleware,
    SkipNonHTMLMiddleware,
    UserAgentMiddleware,
)
from silkworm.pipelines import JsonLinesPipeline


class QuotesSpider(Spider):
    name = "quotes"
    start_urls = ("https://quotes.toscrape.com/",)

    async def parse(self, response: Response):
        if not isinstance(response, HTMLResponse):
            return

        html = response
        for quote in await html.select(".quote"):
            text_el = await quote.select_first(".text")
            author_el = await quote.select_first(".author")
            if text_el is None or author_el is None:
                continue
            tags = await quote.select(".tag")
            yield {
                "text": text_el.text,
                "author": author_el.text,
                "tags": [t.text for t in tags],
            }

        if next_link := await html.select_first("li.next > a"):
            yield html.follow(next_link.attr("href"), callback=self.parse)


if __name__ == "__main__":
    run_spider(
        QuotesSpider,
        request_middlewares=[UserAgentMiddleware()],
        response_middlewares=[
            SkipNonHTMLMiddleware(),
            RetryMiddleware(max_times=3, sleep_http_codes=[429, 503]),
        ],
        item_pipelines=[JsonLinesPipeline("data/quotes.jl")],
        concurrency=16,
        request_timeout=10,
        log_stats_interval=30,
    )
```

`run_spider`/`crawl` knobs:
- `concurrency`: number of concurrent HTTP requests; default 16.
- `max_pending_requests`: queue bound to avoid unbounded memory use (defaults to `concurrency * 10`).
- `request_timeout`: per-request timeout (seconds).
- `keep_alive`: reuse HTTP connections when supported by the underlying client (sends `Connection: keep-alive`).
- `http_client`: use a custom client instance such as `OnionLinkClient(...)` instead of the default wreq-backed client.
- `html_max_size_bytes`: limit HTML parsed into `AsyncDocument` to avoid huge payloads.
- `log_stats_interval`: seconds between periodic stats logs; final stats are always emitted.
- `request_middlewares` / `response_middlewares` / `item_pipelines`: plug-ins run on every request/response/item.
- use `run_spider_rsloop(...)` instead of `run_spider(...)` to run under rsloop (requires `pip install silkworm-rs[rsloop]`).
- use `run_spider_uvloop(...)` instead of `run_spider(...)` to run under uvloop (requires `pip install silkworm-rs[uvloop]`).
- use `run_spider_winloop(...)` instead of `run_spider(...)` to run under winloop on Windows (requires `pip install silkworm-rs[winloop]`).

## Built-in middlewares and pipelines

```python
from silkworm.middlewares import (
    CloudflareCrawlMiddleware,
    DelayMiddleware,
    ProxyMiddleware,
    RetryMiddleware,
    SkipNonHTMLMiddleware,
    UserAgentMiddleware,
)
from silkworm.pipelines import (
    CallbackPipeline,  # invoke a custom callback function on each item
    CSVPipeline,
    JsonLinesPipeline,
    MsgPackPipeline,  # requires: pip install silkworm-rs[msgpack]
    SQLitePipeline,
    XMLPipeline,
    TaskiqPipeline,  # requires: pip install silkworm-rs[taskiq]
    PolarsPipeline,  # requires: pip install silkworm-rs[polars]
    ExcelPipeline,  # requires: pip install silkworm-rs[excel]
    YAMLPipeline,  # requires: pip install silkworm-rs[yaml]
    AvroPipeline,  # requires: pip install silkworm-rs[avro]
    ElasticsearchPipeline,  # requires: pip install silkworm-rs[elasticsearch]
    MongoDBPipeline,  # requires: pip install silkworm-rs[mongodb]
    MySQLPipeline,  # requires: pip install silkworm-rs[mysql]
    PostgreSQLPipeline,  # requires: pip install silkworm-rs[postgresql]
    S3JsonLinesPipeline,  # requires: pip install silkworm-rs[s3]
    VortexPipeline,  # requires: pip install silkworm-rs[vortex]
    WebhookPipeline,  # sends items to webhook endpoints using wreq
    GoogleSheetsPipeline,  # requires: pip install silkworm-rs[gsheets]
    SnowflakePipeline,  # requires: pip install silkworm-rs[snowflake]
    FTPPipeline,  # requires: pip install silkworm-rs[ftp]
    SFTPPipeline,  # requires: pip install silkworm-rs[sftp]
    CassandraPipeline,  # requires: pip install silkworm-rs[cassandra]
    CouchDBPipeline,  # requires: pip install silkworm-rs[couchdb]
    DynamoDBPipeline,  # requires: pip install silkworm-rs[dynamodb]
    DuckDBPipeline,  # requires: pip install silkworm-rs[duckdb]
)

run_spider(
    QuotesSpider,
    request_middlewares=[
        UserAgentMiddleware(),  # rotate/custom user agent
        DelayMiddleware(min_delay=0.3, max_delay=1.2),  # polite throttling
        # ProxyMiddleware with round-robin selection (default)
        # ProxyMiddleware(proxies=["http://user:pass@proxy1:8080", "http://proxy2:8080"]),
        # ProxyMiddleware with random selection
        # ProxyMiddleware(proxies=["http://proxy1:8080", "http://proxy2:8080"], random_selection=True),
        # ProxyMiddleware from file with random selection
        # ProxyMiddleware(proxy_file="proxies.txt", random_selection=True),
    ],
    response_middlewares=[
        RetryMiddleware(max_times=3, sleep_http_codes=[403, 429]),  # backoff + retry
        SkipNonHTMLMiddleware(),  # drop callbacks for images/APIs/etc
    ],
    item_pipelines=[
        JsonLinesPipeline("data/quotes.jl"),
        SQLitePipeline("data/quotes.db", table="quotes"),
        XMLPipeline("data/quotes.xml", root_element="quotes", item_element="quote"),
        CSVPipeline("data/quotes.csv", fieldnames=["author", "text", "tags"]),
        MsgPackPipeline("data/quotes.msgpack"),
    ],
)
```

- `DelayMiddleware` strategies: `delay=1.0` (fixed), `min_delay/max_delay` (random), or `delay_func` (custom).
- `ProxyMiddleware` supports three modes:
  - **Round-robin (default)**: `ProxyMiddleware(proxies=["http://proxy1:8080", "http://proxy2:8080"])` cycles through proxies in order.
  - **Random selection**: `ProxyMiddleware(proxies=["http://proxy1:8080", "http://proxy2:8080"], random_selection=True)` randomly selects a proxy for each request.
  - **From file**: `ProxyMiddleware(proxy_file="proxies.txt")` loads proxies from a file (one proxy per line, blank lines ignored). Combine with `random_selection=True` for random selection from the file.
- `RetryMiddleware` backs off with `asyncio.sleep`; any status in `sleep_http_codes` is retried even if not in `retry_http_codes`.
- `SkipNonHTMLMiddleware` checks `Content-Type` and optionally sniffs the body (`sniff_bytes`) to avoid running HTML callbacks on binary/API responses.
- `CloudflareCrawlMiddleware` is opt-in per request via `request.meta["cloudflare_crawl"]`; it submits a Cloudflare Browser Rendering crawl job, polls until completion, and hands your callback a synthetic JSON `Response` with the final API payload.
- `JsonLinesPipeline` writes items to a local JSON Lines file and, when `opendal` is installed, appends asynchronously via the filesystem backend (`use_opendal=False` to stick to a regular file handle).
- `CSVPipeline` flattens nested dicts (e.g., `{"user": {"name": "Alice"}}` -> `user_name`) and joins lists with commas; `XMLPipeline` preserves nesting.
- `MsgPackPipeline` writes items in binary MessagePack format using [ormsgpack](https://github.com/aviramha/ormsgpack) for fast and compact serialization (requires `pip install silkworm-rs[msgpack]`).
- `TaskiqPipeline` sends items to a [Taskiq](https://taskiq-python.github.io/) queue for distributed processing (requires `pip install silkworm-rs[taskiq]`).
- `PolarsPipeline` writes items to a Parquet file using Polars for efficient columnar storage (requires `pip install silkworm-rs[polars]`).
- `ExcelPipeline` writes items to an Excel .xlsx file (requires `pip install silkworm-rs[excel]`).
- `YAMLPipeline` writes items to a YAML file (requires `pip install silkworm-rs[yaml]`).
- `AvroPipeline` writes items to an Avro file with optional schema (requires `pip install silkworm-rs[avro]`).
- `ElasticsearchPipeline` sends items to an Elasticsearch index (requires `pip install silkworm-rs[elasticsearch]`).
- `MongoDBPipeline` sends items to a MongoDB collection (requires `pip install silkworm-rs[mongodb]`).
- `MySQLPipeline` sends items to a MySQL database table as JSON (requires `pip install silkworm-rs[mysql]`).
- `PostgreSQLPipeline` sends items to a PostgreSQL database table as JSONB (requires `pip install silkworm-rs[postgresql]`).
- `S3JsonLinesPipeline` writes items to AWS S3 in JSON Lines format using async OpenDAL (requires `pip install silkworm-rs[s3]`).
- `VortexPipeline` writes items to a [Vortex](https://github.com/spiraldb/vortex) file for high-performance columnar storage with 100x faster random access and 10-20x faster scans compared to Parquet (requires `pip install silkworm-rs[vortex]`).
- `WebhookPipeline` sends items to webhook endpoints via HTTP POST/PUT using wreq (same HTTP client as the spider) with support for batching and custom headers.
- `GoogleSheetsPipeline` appends items to Google Sheets with automatic flattening of nested data structures (requires `pip install silkworm-rs[gsheets]` and service account credentials).
- `SnowflakePipeline` sends items to Snowflake data warehouse tables as JSON (requires `pip install silkworm-rs[snowflake]`).
- `FTPPipeline` writes items to an FTP server in JSON Lines format (requires `pip install silkworm-rs[ftp]`).
- `SFTPPipeline` writes items to an SFTP server in JSON Lines format with support for password or key-based authentication (requires `pip install silkworm-rs[sftp]`).
- `CassandraPipeline` sends items to Apache Cassandra database tables (requires `pip install silkworm-rs[cassandra]`).
- `CouchDBPipeline` sends items to CouchDB databases as documents (requires `pip install silkworm-rs[couchdb]`).
- `DynamoDBPipeline` sends items to AWS DynamoDB tables with automatic table creation (requires `pip install silkworm-rs[dynamodb]`).
- `DuckDBPipeline` sends items to a DuckDB database table as JSON (requires `pip install silkworm-rs[duckdb]`).
- `CallbackPipeline` invokes a custom callback function (sync or async) on each item, enabling inline processing logic without creating a full pipeline class. See example below.

## Using CallbackPipeline for custom processing
Process items with custom callback functions without creating a full pipeline class:

```python
from silkworm.pipelines import CallbackPipeline

# Sync callback
def print_item(item, spider):
    print(f"[{spider.name}] {item}")
    return item

# Async callback
async def validate_item(item, spider):
    # Could do async operations like database checks
    if len(item.get("text", "")) < 10:
        print(f"Warning: Short text in item")
    return item

# Modifying callback
def enrich_item(item, spider):
    item["spider_name"] = spider.name
    item["processed"] = True
    return item

run_spider(
    QuotesSpider,
    item_pipelines=[
        CallbackPipeline(callback=print_item),
        CallbackPipeline(callback=validate_item),
        CallbackPipeline(callback=enrich_item),
    ],
)
```

Callbacks receive `(item, spider)` and should return the processed item (or `None` to return the original item unchanged).

## Streaming items to a queue with TaskiqPipeline
Stream scraped items to a [Taskiq](https://taskiq-python.github.io/) queue for distributed processing:

```python
from taskiq import InMemoryBroker
from silkworm.pipelines import TaskiqPipeline

broker = InMemoryBroker()

@broker.task
async def process_item(item):
    # Your item processing logic here
    print(f"Processing: {item}")
    # Save to database, send to another service, etc.

pipeline = TaskiqPipeline(broker, task=process_item)
run_spider(MySpider, item_pipelines=[pipeline])
```

This enables distributed processing, retries, rate limiting, and other Taskiq features. See `examples/taskiq_quotes_spider.py` for a complete example.

## Handling non-HTML responses
Keep crawls cheap when URLs mix HTML and binaries/APIs:

```python
response_middlewares=[SkipNonHTMLMiddleware(sniff_bytes=1024)]
# Tighten HTML parsing size (bytes) to avoid loading huge bodies into scraper-rs
run_spider(MySpider, html_max_size_bytes=1_000_000)
```

## Performance optimization with rsloop
For improved async performance, enable rsloop as a drop-in replacement for asyncio's event loop:

```bash
pip install silkworm-rs[rsloop]
# or with uv:
uv pip install silkworm-rs[rsloop]
```

Then call `run_spider_rsloop` (same signature as `run_spider`):

```python
from silkworm import run_spider_rsloop

run_spider_rsloop(
    QuotesSpider,
    concurrency=32,
)
```

## Performance optimization with uvloop
For improved async performance, enable uvloop (a fast, drop-in replacement for asyncio's event loop):

```bash
pip install silkworm-rs[uvloop]
# or with uv:
uv pip install silkworm-rs[uvloop]
```

Then call `run_spider_uvloop` (same signature as `run_spider`):

```python
from silkworm import run_spider_uvloop

run_spider_uvloop(
    QuotesSpider,
    concurrency=32,
)
```

uvloop can provide 2-4x performance improvement for I/O-bound workloads.

## Performance optimization with winloop (Windows)
For Windows users who want improved async performance, enable winloop (a Windows-compatible alternative to uvloop):

```bash
pip install silkworm-rs[winloop]
# or with uv:
uv pip install silkworm-rs[winloop]
```

Then call `run_spider_winloop` (same signature as `run_spider`):

```python
from silkworm import run_spider_winloop

run_spider_winloop(
    QuotesSpider,
    concurrency=32,
)
```

winloop provides significant performance improvements on Windows, similar to what uvloop offers on Unix-like systems.

## Running spiders with trio
If you prefer trio over asyncio, you can use `run_spider_trio` instead of `run_spider`:

```bash
pip install silkworm-rs[trio]
# or with uv:
uv pip install silkworm-rs[trio]
```

Then use `run_spider_trio`:

```python
from silkworm import run_spider_trio

run_spider_trio(
    QuotesSpider,
    concurrency=16,
    request_timeout=10,
)
```

This runs your spider using trio as the async backend via trio-asyncio compatibility layer.

## JavaScript rendering with Lightpanda (CDP)
For pages that require JavaScript execution, you can use Lightpanda (or any CDP-compatible browser) instead of the standard HTTP client. This uses the Chrome DevTools Protocol (CDP) to control a browser.

### Installation
```bash
pip install silkworm-rs[cdp]
# or with uv:
uv pip install silkworm-rs[cdp]
```

### Starting Lightpanda
```bash
lightpanda --remote-debugging-port=9222
```

Or use Chrome/Chromium:
```bash
chromium --remote-debugging-port=9222 --headless
```

### Using CDP in your spider
There are two ways to use CDP: the convenience API or custom spider integration.

#### Convenience API (simple one-off fetches)
```python
import asyncio
from silkworm import fetch_html_cdp

async def main():
    # Fetch HTML with JavaScript rendering
    text, doc = await fetch_html_cdp(
        "https://example.com",
        ws_endpoint="ws://127.0.0.1:9222",
        timeout=30.0
    )
    
    # Extract data from rendered page
    title = doc.select_first("title")
    print(title.text if title else "No title")

asyncio.run(main())
```

#### Full Spider Integration
```python
from silkworm import HTMLResponse, Request, Response, Spider
from silkworm.cdp import CDPClient

class LightpandaSpider(Spider):
    name = "lightpanda"
    start_urls = ("https://example.com/",)

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        self._cdp_client = None

    async def start_requests(self):
        # Connect to CDP endpoint
        self._cdp_client = CDPClient(
            ws_endpoint="ws://127.0.0.1:9222",
            timeout=30.0
        )
        await self._cdp_client.connect()
        
        for url in self.start_urls:
            yield Request(url=url, callback=self.parse)

    async def parse(self, response: Response):
        if not isinstance(response, HTMLResponse):
            return
        
        # Extract links from JavaScript-rendered page
        for link in await response.select("a"):
            href = link.attr("href")
            if href:
                yield {"url": href}

    async def close(self):
        if self._cdp_client:
            await self._cdp_client.close()
```

See `examples/lightpanda_simple.py` and `examples/lightpanda_spider.py` for complete working examples.

**Note:** CDP support is experimental. For production use, consider using dedicated browser automation tools or the standard HTTP client when JavaScript rendering is not required.

## Onion services with OnionLink
For Tor v3 `.onion` sites, install the optional OnionLink extra and pass `OnionLinkClient` as the spider HTTP client:

```bash
pip install "silkworm-rs[onionlink]"
```

```python
from silkworm import HTMLResponse, OnionLinkClient, Response, Spider, run_spider


class OnionSpider(Spider):
    name = "onion"
    start_urls = ("http://exampleexampleexampleexampleexampleexampleexampleexampleexampleexample.onion/",)

    async def parse(self, response: Response):
        if isinstance(response, HTMLResponse):
            title = await response.select_first("title")
            yield {"title": title.text if title else ""}


run_spider(
    OnionSpider,
    http_client=OnionLinkClient(concurrency=4, timeout=30),
)
```

`OnionLinkClient` supports Silkworm `Request` headers, `params`, body/data, JSON payloads, redirects, HTML detection, and `request.meta["redirect_times"]`. Override OnionLink's response byte cap per request with `request.meta["onionlink_response_limit"]`.

## Logging and crawl statistics
- Structured logs via `logly`; set `SILKWORM_LOG_LEVEL=DEBUG` for verbose request/response/middleware output.
- Periodic statistics with `log_stats_interval`; final stats always include elapsed time, queue size, requests/sec, seen URLs, items scraped, errors, and memory MB.

## Limitations
- By default, HTTP fetches are wreq-based without JavaScript execution; pages requiring client-side rendering can use the optional CDP integration (see "JavaScript rendering with Lightpanda" section) or external browser automation tools. Tor v3 `.onion` sites can use the optional OnionLink integration.
- Request deduplication keys only on `Request.url`; query params, HTTP method, and body are ignored, so same-URL requests with different params/data are dropped unless you set `dont_filter=True` or make the URL unique yourself.
- HTML parsing auto-detects encoding (BOM, HTTP headers/meta, charset detection fallback) but still enforces a `html_max_size_bytes`/`doc_max_size_bytes` cap (default 5 MB) in `scraper-rs` selectors, so very large pages may need a higher limit or preprocessing.
- Several pipelines buffer all items in memory until close (PolarsPipeline, ExcelPipeline, YAMLPipeline, AvroPipeline, VortexPipeline, S3JsonLinesPipeline, FTPPipeline, SFTPPipeline), which can bloat RAM on long crawls; prefer streaming pipelines like JsonLines/CSV/SQLite for high-volume runs.
- Many destination pipelines rely on optional extras; CassandraPipeline is disabled on Windows because `cassandra-driver` depends on libev there.

## Examples
- `python examples/quotes_spider.py` → `data/quotes.jl`
- `python examples/quotes_spider_trio.py` → `data/quotes_trio.jl` (demonstrates trio backend)
- `python examples/quotes_spider_winloop.py` → `data/quotes_winloop.jl` (demonstrates winloop backend for Windows)
- `python examples/hackernews_spider.py --pages 5` → `data/hackernews.jl`
- `python examples/lobsters_spider.py --pages 2` → `data/lobsters.jl`
- `python examples/url_titles_spider.py --urls-file data/url_titles.jl --output data/titles.jl` (includes `SkipNonHTMLMiddleware` and stricter HTML size limits)
- `python examples/export_formats_demo.py --pages 2` → JSONL, XML, and CSV outputs in `data/`
- `python examples/taskiq_quotes_spider.py --pages 2` → demonstrates TaskiqPipeline for queue-based processing
- `python examples/sitemap_spider.py --sitemap-url https://example.com/sitemap.xml --pages 50` → `data/sitemap_meta.jl` (extracts meta tags and Open Graph data from sitemap URLs)
- `python examples/lightpanda_simple.py` → demonstrates CDP/Lightpanda for JavaScript rendering (requires `pip install silkworm-rs[cdp]` and running Lightpanda)
- `python examples/lightpanda_spider.py` → full spider example using CDP/Lightpanda

## Convenience API
For one-off fetches without a full spider:

### Standard HTTP fetch
```python
import asyncio
from silkworm import fetch_html

async def main():
    text, doc = await fetch_html("https://example.com")
    title = await doc.select_first("title")
    print(title.text if title else "No title")

asyncio.run(main())
```

### CDP-based fetch (with JavaScript rendering)
```python
import asyncio
from silkworm import fetch_html_cdp

async def main():
    # Requires Lightpanda/Chrome running with CDP enabled
    text, doc = await fetch_html_cdp("https://example.com")
    title = await doc.select_first("title")
    print(title.text if title else "No title")

asyncio.run(main())
```

## Contributing
Pull requests and issues are welcome. To set up a dev environment, install [uv](https://docs.astral.sh/uv/getting-started/), create a Python 3.13 virtualenv, and sync dev dependencies:

```bash
uv venv --python python3.13
uv sync --group dev
```

Run the checks before opening a PR:

```bash
just fmt && just lint && just typecheck && just test
```

## Acknowledgements
Silkworm is built on top of excellent open-source projects:

- [wreq](https://github.com/0x676e67/wreq-python) - HTTP client with browser impersonation capabilities
- [onionlink](https://github.com/RustedBytes/onionlink-rs) - Tor v3 onion-service client
- [scraper-rs](https://github.com/RustedBytes/scraper-rs) - Fast HTML parsing library
- [logly](https://github.com/muhammad-fiaz/logly) - Structured logging
- [rxml](https://github.com/nephi-dev/rxml) - XML parsing and writing

We are grateful to the maintainers and contributors of these projects for their work.

## License
MIT License. See `LICENSE` for details.
