Given a link to filing document (e.g. the 10-K, 8-K) in HTML, process the file into parts and items. This enables follow-up processing of a desired section - e.g. just the Risk Factors. `item.name` and `part.name` are taken directly from the document without any attempt to normalize.

parse_filing(x, strip = TRUE, include.raw = FALSE, fix.errors = TRUE)

Arguments

x

- URL to a filing HTML document, html text or xml_document

strip

- Should non-text elements be removed? Default: true

include.raw

- Include unprocessed nodes in result? Default: false

fix.errors

- Try to fix document errors (e.g. missing part labels). WIP. Default: true

Value

a dataframe with one row per paragraph

part.name

Detected name of the Part

item.name

Detected name of the Item

text

Text of the paragraph / node

raw*

Raw HTML of the node if include.raw = TRUE

Details

NOTE: This has been tested on a range of documents, but formatting differences could cause failures. Please report an issue for any document that isn't parsed correctly. FURTHER NOTE: Not all filings are well formed - missing headings, bad spacing, etc. These can all throw the parsing off!

Examples

head(parse_filing(paste0('https://www.sec.gov/Archives/edgar/data/', '712515/000071251517000010/ea12312016-q3fy1710qdoc.htm')), 6)
#> text part.name item.name #> 1 Document #> 2 Table of Contents #> 3 UNITED STATES #> 4 SECURITIES AND EXCHANGE COMMISSION #> 5 WASHINGTON, D.C. 20549 #> 6 FORM 10-Q