ashtum / boost-gecko

A crawler that extracts search records for boost.org libraries. demo: https://cppalliance.org/boost-gecko/

Boost-Gecko

A crawler with the goal of creating a search index by crawling HTML documents for every library in the Boost release archive.

How it works

It extracts contents with the correct hierarchy for each library and generates records like this:

{
    "type": "content",
    "library_key": "url",
    "library_name": "URL",
    "content": "this->has_fragment() == false && this->encoded_fragment() == \"\"",
    "weight": {
        "pageRank": 0,
        "level": 60,
        "position": 0
    },
    "hierarchy": {
        "lvl0": {
            "title": "Reference",
            "path": "libs/url/doc/html/url/ref.html"
        },
        "lvl1": {
            "title": "url",
            "path": "libs/url/doc/html/url/ref/boost__urls__url.html"
        },
        "lvl2": {
            "title": "url::remove_fragment",
            "path": "libs/url/doc/html/url/ref/boost__urls__url/remove_fragment.html"
        },
        "lvl3": {
            "title": "Postconditions",
            "path": "libs/url/doc/html/url/ref/boost__urls__url/remove_fragment.html#url.ref.boost__urls__url.remove_fragment.postconditions"
        },
        "lvl4": null,
        "lvl5": null,
        "lvl6": null
    }
}

How to use

Configuration

config.yaml contains all the information that Boost-Gecko requires to extract search records and upload them to Algolia.

Tip: You can use config.local.yaml for development purposes (e.g., working on a single crawler); this file has higher priority over config.yaml and is listed in .gitignore.

Downloading and extracting Boost release archive

Assuming we are in boost-gecko directory, we download and extract Boost release archive in ../boost_1_85_0.

wget https://boostorg.jfrog.io/artifactory/main/release/1.85.0/source/boost_1_85_0.tar.gz
tar -xzf boost_1_85_0.tar.gz -C ../

Extracting search records

The following command utilizes the crawler and library list specified in the config to extract search records and store them in the algolia_records directory.

python -m gecko.extract_records # Utilizes all available cores
# or
python -m gecko.extract_records --concurrency=4 # Utilizes 4 cores

Validating the generated records

The purpose of this step is to identify unexpected issues between releases. For example, a library may have changed its documentation method, causing the configured crawler to fail in extracting search records.

The following command validates whether the numbers of records, words, and lvls generated from the previous step are within 5% of the expected numbers in the configuration or not.

Note: It is quite common to encounter false positive errors in this step due to significant changes in library documentation. The objective is to ensure that changes are valid, update the expected numbers in the configuration, and rerun the process.

python -m gecko.sanitizer check

Updating configuration with the generated records

The following command updates config.yaml with the latest numbers of records, words, and lvls for each library. These numbers will be utilized for the validation step in the next release.

python -m gecko.sanitizer update-config

Uploading Algolia records

The following command uploads records from the algolia_records directory. It uses the boost.version from the config file for the index name.

Note: algolia.settings from the config file would be used for the index settings on Aloglia; no need to adjust index settings via dashboard.

python -m gecko.index_on_algolia

Acknowledgments

• Tomasz Kalisiak for adding a crawler for Boost.Hana library.