Search Engine Parser

"If it is a search engine, then it can be parsed" - some random guy

search-engine-parser is a package that lets you query popular search engines and scrape for result titles, links, descriptions and more. It aims to scrape the widest range of search engines. View all supported engines here.

• Search Engine Parser
  • Popular Supported Engines
  • Installation
  • Development
  • Code Documentation
  • Running the tests
  • Usage
    • Code
    • Command line
  • FAQ
  • Code of Conduct
  • Contribution
  • License (MIT)

Popular Supported Engines

Popular search engines supported include:

• Google
• DuckDuckGo
• GitHub
• StackOverflow
• Baidu
• YouTube

View all supported engines here.

Installation

Install from PyPi:

    # install only package dependencies
    pip install search-engine-parser
    # Installs `pysearch` cli  tool
    pip install "search-engine-parser[cli]"

or from master:

  pip install git+https://github.com/bisoncorps/search-engine-parser

Development

Clone the repository:

    git clone [email protected]:bisoncorps/search-engine-parser.git

Then create a virtual environment and install the required packages:

    mkvirtualenv search_engine_parser
    pip install -r requirements/dev.txt

Code Documentation

Code docs can be found on Read the Docs.

Running the tests

    pytest

Usage

Code

Query results can be scraped from popular search engines, as shown in the example snippet below.

  import pprint

  from search_engine_parser.core.engines.bing import Search as BingSearch
  from search_engine_parser.core.engines.google import Search as GoogleSearch
  from search_engine_parser.core.engines.yahoo import Search as YahooSearch

  search_args = ('preaching to the choir', 1)
  gsearch = GoogleSearch()
  ysearch = YahooSearch()
  bsearch = BingSearch()
  gresults = gsearch.search(*search_args)
  yresults = ysearch.search(*search_args)
  bresults = bsearch.search(*search_args)
  a = {
      "Google": gresults,
      "Yahoo": yresults,
      "Bing": bresults
      }

  # pretty print the result from each engine
  for k, v in a.items():
      print(f"-------------{k}------------")
      for result in v:
          pprint.pprint(result)

  # print first title from google search
  print(gresults["titles"][0])
  # print 10th link from yahoo search
  print(yresults["links"][9])
  # print 6th description from bing search
  print(bresults["descriptions"][5])

  # print first result containing links, descriptions and title
  print(gresults[0])

For localization, you can pass the url keyword and a localized url. This queries and parses the localized url using the same engine's parser:

  # Use google.de instead of google.com
  results = gsearch.search(*search_args, url="google.de")

If you need results in a specific language you can pass the 'hl' keyword and the 2-letter country abbreviation (here's a handy list):

  # Use 'it' to receive italian results
  results = gsearch.search(*search_args, hl="it")

Cache

The results are automatically cached for engine searches. You can either bypass the cache by adding cache=False to the search or async_search method or clear the engine's cache

    from search_engine_parser.core.engines.github import Search as GitHub
    github = GitHub()
    # bypass the cache
    github.search("search-engine-parser", cache=False)

    #OR
    # clear cache before search
    github.clear_cache()
    github.search("search-engine-parser")

Proxy

Adding a proxy entails sending details to the search function

    from search_engine_parser.core.engines.github import Search as GitHub
    github = GitHub()
    github.search("search-engine-parser", 
        # http proxies supported only
        proxy='http://123.12.1.0',
        proxy_auth=('username', 'password'))

Async

search-engine-parser supports async:

   results = await gsearch.async_search(*search_args)

Results

The SearchResults after searching:

  >>> results = gsearch.search("preaching to the choir", 1)
  >>> results
  <search_engine_parser.core.base.SearchResult object at 0x7f907426a280>
  # the object supports retrieving individual results by iteration of just by type (links, descriptions, titles)
  >>> results[0] # returns the first <SearchItem>
  >>> results[0]["description"] # gets the description of the first item
  >>> results[0]["link"] # gets the link of the first item
  >>> results["descriptions"] # returns a list of all descriptions from all results

It can be iterated like a normal list to return individual SearchItems.

Command line

search-engine-parser comes with a CLI tool known as pysearch. You can use it as such:

pysearch --engine bing search --query "Preaching to the choir" --type descriptions

Result:

'Preaching to the choir' originated in the USA in the 1970s. It is a variant of the earlier 'preaching to the converted', which dates from England in the late 1800s and has the same meaning. Origin - the full story 'Preaching to the choir' (also sometimes spelled quire) is of US origin.

There is a needed argument for the CLI i.e -e Engine followed by either of two subcommands in the CLI i.e search and summary

usage: pysearch [-h] [-u URL] [-e ENGINE] {search,summary} ...

SearchEngineParser

positional arguments:
  {search,summary}      help for subcommands
    search              search help
    summary             summary help

optional arguments:
  -h, --help            show this help message and exit
  -u URL, --url URL     A custom link to use as base url for search e.g
                        google.de
  -e ENGINE, --engine ENGINE
                        Engine to use for parsing the query e.g google, yahoo,
                        bing, duckduckgo (default: google)

summary returns the summary of the specified search engine:

pysearch --engine google summary

Full arguments for the search subcommand shown below

usage: pysearch search [-h] -q QUERY [-p PAGE] [-t TYPE] [-r RANK]

optional arguments:
  -h, --help            show this help message and exit
  -q QUERY, --query QUERY
                        Query string to search engine for
  -p PAGE, --page PAGE  Page of the result to return details for (default: 1)
  -t TYPE, --type TYPE  Type of detail to return i.e full, links, desciptions
                        or titles (default: full)
  -r RANK, --rank RANK  ID of Detail to return e.g 5 (default: 0)
  -cc, --clear_cache    Clear cache of engine before searching

Code of Conduct

Make sure to adhere to the code of conduct at all times.

Contribution

Before making any contributions, please read the contribution guide.

License (MIT)

This project is licensed under the MIT 2.0 License which allows very broad use for both academic and commercial purposes.

Contributors ✨

Thanks goes to these wonderful people (emoji key):

Ed Luff 💻	Diretnan Domnan 🚇 ⚠️ 🔧 💻	MeNsaaH 🚇 ⚠️ 🔧 💻	Aditya Pal ⚠️ 💻 📖	Avinash Reddy 🐛	David Onuh 💻 ⚠️	Panagiotis Simakis 💻 ⚠️
reiarthur 💻	Ashokkumar TA 💻	Andreas Teuber 💻	mi096684 🐛	devajithvs 💻	Geg Zakaryan 💻 🐛	Hakan Boğan 🐛
NicKoehler 🐛 💻	ChrisLin 🐛 💻	Pietro 💻 🐛

This project follows the all-contributors specification. Contributions of any kind welcome!