Skip to content

Latest commit

 

History

History
221 lines (173 loc) · 8.22 KB

NEWS.md

File metadata and controls

221 lines (173 loc) · 8.22 KB

edgarWebR 1.0.3.9999

Features

  • Support new EDGAR full-text search
  • Add user-agent setting

Internal

  • Change to use with_mock_api from httptest

edgarWebR 1.0.3

Bugfixes

  • Tests using dev testdata are skipped on cran to reduce overall package size

Internal

  • Add github actions for testing to expand and normalize testing

edgarWebR 1.0.2

Bugfixes

  • User options = 'HUGE' for all xml/html parsing - should resolve some test failures on Solaris.
  • All vignettes now use cached requests (tests were all ready using caches)

Internal

  • httptest redaction/resolving logic centralized in a more standard way

edgarWebR 1.0.1

Bugfixes

  • parse_text_filing() correctly parses filings that have spaces in separating lines. (Fixes #4)
  • parse_text_filing() correctly removes and splits on '' dividers, including when there is no page number. (fixes #6)
  • Initial fix for #11 where some documents with especially deep structures would error in the parsing phase.
  • Check for errors connecting to SEC servers and stop on failures.

Documentation

  • Elaborate the parsing vignette to better explain the files processed by parse_filing() and parse_submission()
  • Document output of parse_submission()
  • Document how the SEC Company 'Fast Search' works and how edgarWebR implements it. (addresses #7)

Internal

  • Add missing pkgdown dev dependency
  • Add lintr dev dependency
  • Update tests w/ current data

edgarWebR 1.0.0 - "Everything in its place"

Features

  • sic_codes a dataset for sic codes from US Department of Labor
  • parse_submission() has options for skipping collection of all content or just binary content
  • Added utility functions submission_index_href(), submission_href(), and submission_file_href() for creating links to filings and their components.
  • Added cik_search() to lookup CIK codes for companies.
  • Added current_events() to access recent filings by form type
  • Added company_search()
  • Added variable_insurance_search() and variable_insurance_fast_search()
  • Added fund_fast_search()
  • Added effectiveness()

Bugfixes

  • parse_filing() properly processes recent AIG & Costco filings
  • parse_filing() properly processes ADM filings for fy08 & fy09
  • parse_filing() now fails gracefully given a unparsable filing
  • parse_filing() properly handles extraneous '' tags from ~08/09 filings such as ADP
  • parse_filing() handles more edge-cases and has similar word counts to plain document.
  • parse_submission() handles large files successfully using temp files.
  • parse_submission() handles submissions with single files correctly.
  • parse_submission() documented properly.
  • full_text() escapes the search query properly.

Internal

  • Testing now requires tokenizers for testing if parsing adds words (it often does...)
  • Moved to testdata for parse_* functions rather than http cache
  • Made wordcount comparison in test_parse_filing far more restrictive (good thing!)
  • Tried to make TravisCI work, but still not great...
  • Added series_search to cover the main work for fund and variable insurance searches.

edgarWebR 0.3.1 - "Oh yeah, data changes"

Bugfixes

  • full_text no longer tests example as it often is long-running
  • header_search no longer tests example as it often is long-running
  • Vignette 'parsing' now always fetches the same filing to avoid issues with newer filings changing the output

edgarWebR 0.3.0 - "Going all the way back"

Features

  • full_text provides access to the full-text filing search interface.
  • latest_filings provides access to the latest SEC filings.
  • parse_text_filing parses text-only 10-* filings.
  • parse_filing detects when a filing is HTML wrapped plain text and uses parse_text_filing when appropriate.
  • header_search provides access to search filing headers back to 1994

Bugfixes

  • parse_filing now treats <br> as a space avoiding words separated only by a line return getting concatenated. (Fixes #2)
  • parse_filing now tries to detect and remove TOC itemss/parts to avoid duplicate entries

Internal

  • map_xml processes href's out of javascript links
  • map_xml add parameter for date format
  • linting setup to run in CI

edgarWebR 0.2.1 - "Thanks Testers"

Many thanks to everyone providing feedback, particularly Günter Leitold for rigorously testing the parse_filing funcion.

Bugfixes

  • Parser covers a wider range of document formats for 10-K's.
  • filing_documents correctly provides the href for the document not the iXBRL viewer when present.

Internal

  • Limited testing of parse_filing on CRAN.

edgarWebR 0.2.0 - "Parse all the things"

Features

Parse Filings

New functionality in parse_filing to annotate a filing with part and item identifiers. This enables easy extraction of risks or management discussions.

NOTE: While tested on a range of documents, due to the high variation in filing formats, it is likely not to work on every document. Please open an issue if errors are encountered.

Parse Submission

New function parse_submission extracts documents from the full submission file. This is needed for older submissions and accessing documents that are not extracted individually online.

Internal

  • Incorporated httptest for cached testing. Altered most functions to use httr for remote calls to support this test method.
  • New Makefile targets to support cached testing -
    • make test - uses cached responses
    • make test-nocache - makes all remote calls
    • make test-buildcache - runs tests, caching remote requests.
    • make test-cleancache - deletes the test cache

edgarWebR 0.1.1 - "Whoops 1"

Bugfixes

  • Removed a test for the "changed_date" field in filing information as the source test document removed the field. Permanent fix will be to use a a more stable data source for testing.

edgarWebR 0.1.0 - "Hello World"

First CRAN release of edgarWebR. At this point, function parameters, return colums and general shape of data will not change without a deprecation process and due warning.

Features

Searching

  • The company_information(), company_details(), and company_filings() for searching and getting information for a given company or filer.
  • fund_search() searches for mutual funds and fund families.

Filings

  • filing_information(), filing_documents(), filing_filers(), and filing_funds() extract information sections from a particular filing.
  • filing_details() extracts all 4 components from a particular filing.

edgarWebR 0.0.3 - "Now Getting it Ready"

UNSTABLE - Function calls and returns likely to change without warning

Features

  • All methods which took a URL as a parameter now accept either a href or pre-loaded xml document. This facilitates loading a filing from a local file, fetching the filing separately for customized parsing, and testing.
  • Some column names have shifted to better align to overall naming scheme

Package

  • [vignettes] changed Introduction to use purrr for clearer code

Internal

  • Added caches and markdown folders to ignore files
  • Made map_xml handle parsing integers
  • Added live_test to Makefile
  • Way more tests
  • coverage target added to makefile

edgarWebR 0.0.2 - "XPath is is powerful"

UNSTABLE - Function calls and returns likely to change without warning

Features

  • [filings] filing_information provides general filing information
  • [filings] filing_funds gives all funds associated with a filing
  • [filings] filing_filers gives all filers associated with a filing
  • [filings] filing_details returns the information, funds, filers and documents for a filing

Package

  • [vignettes] We now have a basic introductory vignette
  • [tests] initial testes created

Internal

  • [map_xml] map_xml now parses date/time columns into POSIXlt dates
  • [browse_edgar] escape text fields

edgarWebR 0.0.1 - Initial Release

UNSTABLE - Function calls and returns likely to change without warning