uniparser 0.0.6

Provide a universal solution for crawler platforms. Read more: https://github.com/ClericPy/uniparser.

uniparser

Provide a universal solution for crawler, Python3.6+.

Install

pip install uniparser -U

Why?

1. Reduced the code quantity from plenty of similar crawlers & parsers. Don't Repeat Yourself.
2. Make the parsing process of different parsers persistent.
3. Separating parsing processes from the downloading.
4. Provide a universal solution for crawler platforms.
5. Summarize common string parsing tools on the market.

Feature List

1. Support most of popular parsers for HTML / XML / JSON / AnyString / Python object
   1. css (HTML)
   2. xml
   3. regex
   4. jsonpath
   5. objectpath
   6. jmespath
   7. time
   8. loader
      1. json / yaml / toml
   9. waiting for new ones...
2. Request args persistence, support curl-string, single-url, dict, json.
3. A simple Web UI for generate & test CrawlerRule.
4. Serializable JSON rule class for saving the whole parsing process.
   1. Each ParseRule / CrawlerRule / HostRules subclass can be json.dumps to JSON for persistence.
   2. Therefore, they also can be loaded from JSON string.
   3. Nest relation of rule names will be treat as the result format. (Rule's result will be ignore if has childs.)

Quick Start

# -*- coding: utf-8 -*-
from urllib.parse import urlparse

import requests
from uniparser import CrawlerRule, HostRules, Uniparser


def test_default_usage():
    # 1. prepare for storage to save {'host': HostRules}
    uni = Uniparser()
    storage = {}
    test_url = 'http://httpbin.org/get'
    crawler_rule = CrawlerRule(
        'test_crawler_rule',
        {
            'url': 'http://httpbin.org/get',
            'method': 'get'
        },
        [{
            "name": "rule1",
            "chain_rules": [
                ['objectpath', 'JSON.url', ''],
                ['python', 'getitem', '[:4]'],
                ['udf', '(context.url, input_object)', ''],
            ],
            "child_rules": []
        }],
        'https?://httpbin.org/get',
    )
    host = urlparse(test_url).netloc
    hrs = HostRules(host=host)
    hrs.add(crawler_rule)
    # same as: json_string = hrs.to_json()
    json_string = hrs.dumps()
    # print(json_string)
    assert json_string == r'{"host": "httpbin.org", "crawler_rules": [{"name": "test_crawler_rule", "parse_rules": [{"name": "rule1", "chain_rules": [["objectpath", "JSON.url", ""], ["python", "getitem", "[:4]"], ["udf", "(context.url, input_object)", ""]], "child_rules": []}], "request_args": {"url": "http://httpbin.org/get", "method": "get"}, "regex": "https?://httpbin.org/get"}]}'
    # 2. add HostRules to storage, sometimes save on redis
    storage[hrs['host']] = json_string
    # ============================================
    # start to crawl
    # 1. set a example url
    test_url1 = test_url
    # 2. find the HostRules
    json_string = storage.get(host)
    # 3. HostRules init: load from json
    # same as: hrs = HostRules.from_json(json_string)
    hrs = HostRules.loads(json_string)
    # print(crawler_rule)
    # 4. now search / match the url with existing rules
    crawler_rule = hrs.search(test_url1)
    # print(crawler_rule)
    assert crawler_rule == {
        'name': 'test_crawler_rule',
        'parse_rules': [{
            'name': 'rule1',
            'chain_rules': [['objectpath', 'JSON.url', ''],
                            ['python', 'getitem', '[:4]'],
                            ['udf', '(context.url, input_object)', '']],
            'child_rules': []
        }],
        'request_args': {
            'url': 'http://httpbin.org/get',
            'method': 'get'
        },
        'regex': 'https?://httpbin.org/get'
    }
    # print(hrs.match(test_url1))
    assert crawler_rule == hrs.match(test_url1)
    # 5. send request as crawler_rule's request_args, download the page source code
    resp = requests.request(**crawler_rule['request_args'])
    source_code = resp.text
    # 6. parse the whole crawler_rule as crawler_rule's with uniparser. set context with resp
    assert isinstance(crawler_rule, CrawlerRule)
    result = uni.parse(source_code, crawler_rule, context=resp)
    # print(result)
    assert result == {
        'test_crawler_rule': {
            'rule1': ('http://httpbin.org/get', 'http')
        }
    }
    # ===================== while search failed =====================
    # given a url not matched the pattern
    test_url2 = 'http://notmatch.com'
    crawler_rule2 = hrs.search(test_url2)
    assert crawler_rule2 is None
    # ===================== shared context =====================
    # !!! use context by updating rule.context variable
    new_parse = '''
def parse(input_object):
    context['new_key'] = 'cleared'
    return (input_object, context)
    '''
    crawler_rule.context.update({'new_key': 'new_value'})
    crawler_rule.clear_parse_rules()
    crawler_rule.add_parse_rule({
        'name': 'rule1',
        'chain_rules': [['objectpath', 'JSON.url', ''],
                        ['python', 'getitem', '[:4]'], ['udf', new_parse, '']],
        'child_rules': []
    })
    result = uni.parse(source_code, crawler_rule)
    # print(result)
    assert result == {
        'test_crawler_rule': {
            'rule1': ('http', {
                'new_key': 'cleared'
            })
        }
    }
    # print(crawler_rule.context)
    # now the crawler_rule.context has been updated.
    assert crawler_rule.context == {'new_key': 'cleared'}


if __name__ == "__main__":
    test_default_usage()

Uniparser Test Console Demo (Web UI)

1. Prepare Environment

1. pip install bottle uniparser
2. python -m uniparser 8080

2. open browser http://127.0.0.1:8080/

2.1 Start page

2.2 Prepare the rules

2.3 Read the parse result

Show result as repr(result)

{'HelloWorld': {'rule1-get-first-p': 'Customer name: ', 'rule2-get-legends': [' Pizza Size ', ' Pizza Toppings ']}}

As you see, Crawler Rule's name is the root key, and ParseRule's name as the others.

More Usage

Talk is cheap, code is doc. Poor time to write...

Watch the examples: test_parsers.py

TODO

• Release to pypi.org
  • Upload dist with Web UI
• Add github actions for testing package
• Web UI for testing rules
• Complete the whole doc