Elixir/HTML dump scraper: Difference between revisions
Explain references more |
distracting caption to help background the decorative image |
||
| (7 intermediate revisions by 2 users not shown) | |||
| Line 1: | Line 1: | ||
[[File:1.Magukbaforduló ikerszelvényesek 72dpi.jpg|thumb]] | [[File:1.Magukbaforduló ikerszelvényesek 72dpi.jpg|thumb|''Introverted Millipede'']] | ||
{{Project|status=(beta)|url=https://gitlab.com/wmde/technical-wishes/scrape-wiki-html-dump}} | |||
HTML dumps | A new and wondrous data source has become available to Wikimedia researchers and hobbyists: semantic [https://dumps.wikimedia.org/other/enterprise_html/ HTML dumps] of all articles on wiki. Previously, archived wiki content was only available as raw wikitext, which is notoriously difficult to parse for information, or even to render. MediaWiki's wikitext depends on a number of extensions and is usually tied to a specific site's configuration and user-generated templates. This recursively parsed content is effectively impossible to expand exactly as it appeared at the time it was written, once the templates and software it depends on have drifted. | ||
HTML dumps are an improvement in every way: content, structure and information are expanded, frozen and made available in a form that can be read by ordinary tools—and the original wikitext is still available as RDFa annotations which makes the new format something like a superset. | |||
At my day job doing [[metawikimedia:WMDE_Technical_Wishes|Technical Wishes]] for [https://www.wikimedia.de/ Wikimedia Germany], we found a reason to dive into these new dumps: it's the only reliable way to count the footnotes on each article. I'll go into some detail about why other data sources wouldn't have sufficed, and also why we're counting footnotes in the first place. | |||
== Reference parsing == | == Reference parsing == | ||
=== What are references? === | === What are references? === | ||
References are the little footnotes all over Wikipedia articles:<ref group="example-footnote">This is a footnote body.</ref> These footnotes ground the writing in sources, and are a | References are the little footnotes all over Wikipedia articles:<ref group="example-footnote">This is a footnote body. Often you will see a citation to a book or other source here.</ref> These footnotes ground the writing in sources, and are a distinctive aspect of the Wikipedias' intellectual cultures. | ||
<references group="example-footnote" /> | <references group="example-footnote" /> | ||
=== Why are we counting references? === | === Why are we counting references? === | ||
The Wikimedia Germany Technical Wishes team has taken the past few months to [[metawikimedia:WMDE_Technical_Wishes/Reusing_references|focus on how references are reused]] on wikis. We have some ideas about what needs to be fixed (unfortunately this project is currently on hold), but first we needed to take measurements of the baseline situation in order to better understand how references are used, reused, and | The Wikimedia Germany Technical Wishes team has taken the past few months to [[metawikimedia:WMDE_Technical_Wishes/Reusing_references|focus on how references are reused]] on wikis. We have some ideas about what needs to be fixed (unfortunately this project is currently on hold), but first we needed to take measurements of the baseline situation in order to better understand how references are used, reused, and to evaluate whether our potential intervention is beneficial. | ||
Previous [[metawikimedia:Research:Characterizing_Wikipedia_Citation_Usage|research into citations]] has also measured the HTML-formatted articles, but HTML dumps weren't available at the time so this was accomplished by downloading | Previous [[metawikimedia:Research:Characterizing_Wikipedia_Citation_Usage|research into citations]] has also measured references by starting with the HTML-formatted articles, but HTML dumps weren't available at the time so this was accomplished by downloading each article individually. | ||
For those interested in the | For those interested in the preliminary output of the scraper run, please skip ahead to the [https://phabricator.wikimedia.org/T332032#9011167 raw summary results]. Much detailed statistics for each wiki page and template will be published once we figure out longer-term hosting for the data. | ||
=== | === Obstacles to finding references in wikitext === | ||
A raw reference is straightforward in wikitext and looks like: <code><nowiki><ref>This footnote.</ref></nowiki></code>. If this were the end of the story, it would be simple to parse references. What makes it more complicated is that many references are produced using reusable templates, for example: <code><nowiki>{{sfn|Hacker|Grimwood|2011|p=290}}</nowiki></code>. | A raw reference is straightforward in wikitext and looks like: <code><nowiki><ref>This footnote.</ref></nowiki></code>. If this were the end of the story, it would be simple to parse references. What makes it more complicated is that many references are produced using reusable templates, for example: <code><nowiki>{{sfn|Hacker|Grimwood|2011|p=290}}</nowiki></code>. | ||
If "<nowiki>{{sfn}}</nowiki>" were the only template then we could | If "<nowiki>{{sfn}}</nowiki>" were the only template used to produce references then we could search for "<nowiki><ref>" tags and "{{sfn}}</nowiki>" templates in wikitext. But a [https://en.wikipedia.org/w/index.php?search=insource%3A%3Cref%3E+insource%3A%2F%5B%3C%5Dref%5B%3E%5D%2F+-insource%3A%2Finclude+%5B%3C%5Dref%5B%3E%5D%2F&title=Special:Search&profile=advanced&fulltext=1&ns10=1 search for reference-producing templates] unveils over 12,000 different ref-producing templates on English Wikipedia alone, and these are unique to every other wiki and language edition. | ||
=== References are simple in HTML === | === References are simple in HTML === | ||
| Line 41: | Line 41: | ||
The processing is broken down into small units which each write a single file. If a file already exists, then we skip the corresponding calculation. This general caching technique is known as [[w:Memoization|memoization]]. | The processing is broken down into small units which each write a single file. If a file already exists, then we skip the corresponding calculation. This general caching technique is known as [[w:Memoization|memoization]]. | ||
During each | During each unit file, also write to a checkpoint file at multiples of 100 articles processed. The output file is written in chunks immediately after writing to the checkpoint file, which reduces the window of time in which a crash can result in inconsistency. When resuming, if a checkpoint file is present then the job will skip articles without processing them, until the total count catches up with the count given in the checkpoint file. The overall behavior of the job is therefore "at least once" processing, meaning that the only type of irregularity that can happen is to duplicate work for some articles. | ||
== Concurrency == | == Concurrency == | ||
| Line 49: | Line 49: | ||
The next obvious fork point would be in the phase which makes external API requests, but this is trickier because we want to limit total concurrency across all wikis as well, to avoid overwhelming the service. This should be implemented with a connection pool, ideally one which reuses a small number of connections according to HTTP/1.1 . | The next obvious fork point would be in the phase which makes external API requests, but this is trickier because we want to limit total concurrency across all wikis as well, to avoid overwhelming the service. This should be implemented with a connection pool, ideally one which reuses a small number of connections according to HTTP/1.1 . | ||
== Modularity == | |||
It will be no surprise that the analyses are run as separate units under a pluggable architecture, so that the tool can be reused for various tasks. The callbacks are crude for now and the abstraction is leaky, but it at least accomplishes code encapsulation and easily [https://gitlab.com/wmde/technical-wishes/scrape-wiki-html-dump/-/blob/main/config/config.exs configurable] composition. | |||
Modules must be written in Elixir, but we're also considering a language-agnostic callback if the need arises. | |||
Some aspects of modularity were no fun so we ignored them. For example, [https://gitlab.com/wmde/technical-wishes/scrape-wiki-html-dump/-/blob/main/metrics.md each metric is documented] in one big flat file. You'll also encounter some mild hardcoding such as an entire extra processing phase to make external API requests for parsed map data. | |||