Skip to content

washingtonpost/FastFEC

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

FastFEC

A C program to stream and parse Federal Election Commission (FEC) filings, writing output to CSV.

Installation

Download the latest release and place it on your path, or if you have Homebrew and are on Mac or Linux, you can install via:

brew install fastfec

You can also build a binary yourself following the development instructions below.

Usage

Once FastFEC has been installed, you can run the program by calling fastfec in your terminal:

Usage: fastfec [flags] <id or file> [output directory=output] [override id]
  • [flags]: optional flags which must come before other args; see below
  • <file or id> is either
    • a file, in which case the filing is read from disk at the specified local path
    • a numeric ID (only works with --print-url): prints the possible URLs the filing lives on the FEC docquery website
  • [output directory] is the folder in which CSV files will be written. By default, it is output/.
  • [override id] is an ID to use as the filing ID. If not specified, this ID is pulled out of the first parameter as a numeric component that can be found at the end of the path.

The CLI will read the specified filing from disk and then write output CSVs for each form type in the output directory. The paths of the outputted files are:

  • {output directory}/{filing id}/{form type}.csv

You can also pipe the output of another command in by following this usage:

[some command] | fastfec [flags] <id> [output directory=output]

Flags

The CLI supports the following flags:

  • --include-filing-id / -i: if this flag is passed, then the generated output will include a column at the beginning of every generated file called filing_id that gets passed the filing ID. This can be useful for bulk uploading CSVs into a database
  • --silent / -s : suppress all non-error output messages
  • --warn / -w : show warning messages (e.g. for rows with unexpected numbers of fields or field types that don't match exactly)
  • --no-stdin / -x: disable receiving piped input from other programs (stdin)
  • --print-url / -p: print URLs from docquery.fec.gov (cannot be specified with other flags)

The short form of flags can be combined, e.g. -is would include filing IDs and suppress output.

Examples

Parsing a local filing

fastfec -s 13360.fec fastfec_output/

  • This will run FastFEC in silent mode, parse the local filing 13360.fec, and store the output in CSV files at fastfec_output/13360/.

Downloading and parsing a filing

Get the FEC filing URL needed:

fastfec -p 13360

If you have curl installed, you can then run this command:

curl https://docquery.fec.gov/dcdev/posted/13360.fec | fastfec 13360
  • This will download the filing with ID 13360 from the FEC's servers and stream/parse it, storing the output in CSV files at output/13360/

If you don't have curl installed, you can also download the filing from the URL (https://docquery.fec.gov/dcdev/posted/13360.fec), save the file, and run (is equivalent to the above):

fastfec 13360.fec

Benchmarks

The following was performed on an M1 Macbook Air:

Filing Size Time Memory usage CPU usage
1464847.fec 8.4gb 1m 42s 1.7mb 98%

Local development

Build system

Zig is used to build and compile the project. Download and install the latest version of Zig (>=0.11.0) by following the instructions on the website (you can verify it's working by typing zig in the terminal and seeing help commands).

Dependencies

FastFEC has no external C dependencies. PCRE is bundled with the library to ensure compatibility with Zig's build system and cross-platform compilation.

Building

From the root directory of the repo, run:

zig build
  • The above commands will output a binary at zig-out/bin/fastfec and a shared library file in the zig-out/lib/ directory
  • If you want to only build the library, you can pass -Dlib-only=true as a build option following zig build
  • You can also compile for other operating systems via -Dtarget=x86_64-windows (see here for additional targets)

Testing

Currently, there's C tests for specific parsing/buffer/write/CLI functionality and Python integration tests.

  • Running the C tests: zig build test
  • Running the Python tests:
    cd python
    pip install -r requirements-dev.txt
    tox -e py

See the GitHub test workflow for more info

Scripts

python scripts/generate_mappings.py: A Python script to auto-generate C header files containing column header and type mappings