25f7214830
* fixing typo * keeping same convention * avoid wrong use of word crawling * removing superfluous comment * make the documentation more explicit to prevent error in the key usage, introduces feedback from algolia/docsearch-scraper#438, give instruction about how to use the headless chrome * add synonym, resolves #640 * precise sitemap * fix typos * removing unavailable anchor * fixing wrong URL * resolves #518 by removing unwanted char from anchors * fixing linter issues * fix remark-lint warning
1.9 KiB
1.9 KiB
| layout | title |
|---|---|
| two-columns | How do we build a DocSearch index? |
In this section you will learn how we build a DocSearch index from your page.
Everything starts from your page
{mt-2}
We extract the payload thanks to your set of selectors
{mt-2}
We will focus on the highlighted information depending on your selectors.
We iterate through the HTML flow and build the payload
{mt-2}
This payload will be the only data extracted from your page.
We iterate through the payload and start pushing records
{mt-2}
We index the temporary record when we add an element to it (if
min_indexed_level is set to 0)
We pile up the elements based on the current temporary record
{mt-2}
Base on the position within the flow, we nest elements as much as possible to keep the context and incerease the relevancy.
We iterate until we match a text element
{mt-2}
We override the text element when we find a newer one
{mt-2}
We remove the stashed, deeper elements when we add a higher level
{mt-2}
Contextual information and hierarchy must be updated once we encounter a new level. We are doing that because it highlights a new sub-section not related to the previous one.
If you need any further information, please do not hesitate, send us your feedback.