Powerful easy to use library for accessing edgar filings
Project description
Table of Contents
About the project
edgartools
is a library for working Edgar filings in analytic workflows.
Demo
Get the Common Shares Issued amount from Snowflake's latest 10-Q filing
(Company.for_ticker("SNOW")
.get_filings(form="10-Q")
.latest()
.xbrl()
.to_duckdb().execute(
"""select fact, value, units, end_date from facts
where fact = 'CommonStockSharesIssued'
order by end_date desc limit 1
"""
).df()
)
This example shows what can be done with edgartools.
Under the hood the code does the following
- Use the ticker "SNOW" to get the company's cik from the Company Tickers JSON
- From the cik get the company's filings from the submissions endpoint
https://data.sec.gov/submissions/CIK{cik:010}.json
- Select the latest 10-Q filing
- Download the XBRL file for that filing
- Convert the XBRL data into a pandas dataframe
- Register the dataframe as a DuckDB table
- Execute the SQL and convert to a dataframe
You might not want to chain the operations like this, and strictly speaking it might not be the most efficient, given how much work happens within those lines of code. This guide will show you step by step how to easily get SEC filing data and text into your analytic workflows.
Features
- Download listings of Edgar filing by year, quarter since 1994
- Select an individual filing and download the html, XML or content of any attached file
- View a filing XBRL as a dataframe and query it with SQL
- Search for company by ticker or CIK
- Get a company's filings
- Get a dataset of company's facts e.g. CommonSharesOutstanding
- Query a company's facts as SQL using an in-memory DuckDB database
Installation
pip install edgartools
Usage
Set your Edgar user identity
Before you can access the SEC Edgar API you need to set the identity that you will use to access Edgar. This is usually your name and email, or a company name and email.
Sample Company Name AdminContact@<sample company domain>.com
The user identity is sent in the User-Agent string and the Edgar API will refuse to respond to your request without it.
EdgarTools will look for an environment variable called EDGAR_IDENTITY
and use that in each request.
So, you need to set this environment variable before using it.
Setting EDGAR_IDENTITY in Linux/Mac
export EDGAR_IDENTITY="Michael Mccallum mcalum@gmail.com"
Setting EDGAR_IDENTITY in Windows Powershell
$Env:EDGAR_IDENTITY="Michael Mccallum mcalum@gmail.com"
Alternatively, you can call set_identity
which does the same thing.
from edgar import set_identity
set_identity("Michael Mccallum mcalum@gmail.com")
For more detail see https://www.sec.gov/os/accessing-edgar-data
Using the Company API
With the company API you find a company using the cik or ticker. From the company you can access all their historical filings, and a dataset of the company facts. The SEC's company API also supplies a lot more details about a company including industry, the SEC filer type, the mailing and business address and much more.
Find a company using the cik
The cik is the id that uniquely identifies a company at the SEC. It is a number, but is sometimes shown in SEC Edgar resources as a string padded with leading zero's. For the edgar client API, just use the numbers and omit the leading zeroes.
company = Company.for_cik(1318605)
Find a company using ticker
You can get a company using a ticker e.g. SNOW. This will do a lookup for the company cik using the ticker, then load the company using the cik. This makes it two calls versus one for the cik company lookup, but is sometimes more convenient since tickers are easier to remember that ciks.
Note that some companies have multiple tickers, so you technically cannot get SEC filings for a ticker. You instead get the SEC filings for the company to which the ticker belongs.
The ticker is case-insensitive so you can use Company.for_ticker("snow")
or Company.for_ticker("SNOW")
snow = Company.for_ticker("snow")
Company.for_cik(1832950)
Get filings for a company
To get the company's filings use get_filings()
. This gets all the company's filings that are available from the Edgar submissions endpoint.
company.get_filings()
Filtering filings
You can filter the company filings using a number of different parameters.
class CompanyFilings:
...
def get_filings(self,
*,
form: str | List = None,
accession_number: str | List = None,
file_number: str | List = None,
is_xbrl: bool = None,
is_inline_xbrl: bool = None
):
"""
Get the company's filings and optionally filter by multiple criteria
:param form: The form as a string e.g. '10-K' or List of strings ['10-Q', '10-K']
:param accession_number: The accession number that uniquely identifies an SEC filing e.g. 0001640147-22-000100
:param file_number: The file number e.g. 001-39504
:param is_xbrl: Whether the filing is xbrl
:param is_inline_xbrl: Whether the filing is inline_xbrl
:return: The CompanyFiling instance with the filings that match the filters
"""
The CompanyFilings class
The result of get_filings()
is a CompanyFilings
class. This class contains a pyarrow table with the filings
and provides convenient functions for working with filings.
You can access the underlying pyarrow Table
using the .data
property
filings = company.get_filings()
# Get the underlying Table
data: pa.Table = filings.data
Get a filing by index
To access a filing in the CompanyFilings use the bracket []
notation e.g. filings[2]
filings[2]
Get the latest filing
The CompanyFilings
class has a latest
function that will return the latest Filing
.
So, to get the latest 10-Q filing, you do the following
# Latest filing makes sense if you filter by form type e.g. 10-Q
snow_10Qs = snow.get_filings(form='10-Q')
latest_10Q = snow_10Qs.latest()
# Or chain the function calls
snow.get_filings(form='10-Q').latest()
Get company facts
Facts are an interesting and important dataset about a company accumlated from data the company provides to the SEC.
Company facts are available for a company on the Company Factsf"https://data.sec.gov/api/xbrl/companyfacts/CIK{cik:010}.json"
It is a JSON endpoint and edgartools
parses the JSON into a structured dataset - a pyarrow.Table
.
Getting facts for a company
To get company facts, first get the company, then call company.get_facts()
company = Company.for_ticker("SNOW")
company_facts = company.get_facts()
The result is a CompanyFacts
object which wraps the underlying facts and provides convenient ways of working
with the facts data. To get access to the underyling data use the facts
property.
You can get the facts as a pandas dataframe by calling to_pandas
df = company_facts.to_pandas()
Facts differ among companies. To see what facts are available you can use the facts_meta
property.
Getting the facts as a DuckDB table
Ypu can convert the facts to a DuckDB database which allows you to query the facts using SQL.
company_facts: CompanyFacts = get_company_facts(1318605)
db = company_facts.to_duckdb()
df = db.execute("""
select * from facts
""").df()
Working with a Filing
Once you have a filing you can do many things with it including getting the html text of the filing, get xbrl or xml, or list all the files in the filing.
Getting the html text of a filing
html = filing.html()
To get the html text of the filing call filing.html()
Get the Homepage Url
filing.homepage_url
returns the homepage url of the filing. This is the main index page which lists
all the files attached in the filing
Get the filing homepage
To get access to all the documents on the filing you would call filing.get_homepage()
.
This gives you access to the FilingHomepage
class that you can use to list all the documents
and datafiles on the filing.
Working with XBRL filings
Some filings are in XBRL (eXtensible Business Markup Language) format. These are mainly the newer filings, as the SEC has started requiring this for newer filings.
If a filing is in XBRL format then it opens up a lot more ways to get structured data about that specific filing and also about the company referred to in that filing.
The Filing
class has an xbrl
function that will download, parse and structure the filing's XBRL document if one exists.
If it does not exist, then filing.xbrl()
will return None
.
The function filing.xbrl()
returns a FilingXbrl
instance, which wraps the data, and provides convenient
ways of working with the xbrl data.
filing_xbrl = filing.xbrl()
Using the Filings API
The Filings API allows you to get the Edgar filing indexes published by the SEC. You would use it to get a bulk dataset of SEC filings for a given time period. With this dataset, you could filter by form type, by date or by company, though if you intend to filter by a singe company, you should use the Company API.
The get_filings function
The main way to use the Filings API is by get_filings
get_filings
accepts the following parameters
- year a year
2015
, a List of years[2013, 2015]
or arange(2013, 2016)
- quarter a quarter
2
, a List of quarters[1,2,3]
or arange(1,3)
- index this is the type of index. By default it is
"form"
. If you want only XBRL filings use"xbrl"
. You can also use"company"
but this will give you the same dataset as"form"
, sorted by company instead of by form
Get filings for 2021
filings = get_filings(2021)
Get filings for 2021 quarter 4
Instead of getting the filings for an entire year, you can get the filings for a quarter.
filings = get_filings(2021, 4)
Get filings between 2010 and 2019
You can get the filings for a range of years, since the year
parameter accepts a value, a list or a range.
filings = get_filings(range(2010, 2020))
Get XBRL filings for 2022
filings = get_filings(2022, index="xbrl")
Filtering Filings by form
You can filter by form type by providing a form or list of forms.
filings = get_filings(2022, form="10-K")
# Filter by list of forms
filings = get_filings(2022, form=["10-K", "10-Q"])
This will include form amendments e.g. "10-K/A" and "10-Q/A". To not include these set amendments=False
# Filter by list of forms not including amendments
filings = get_filings(2022, form=["10-K", "10-Q"], amendments=False)
The Filings class
The get_filings
returns a Filings
class, which wraps the data returned and provide convenient ways for working with filings.
Convert the filings to a pandas dataframe
The filings data is stored in the Filings
class as a pyarrow.Table
. You can get the data as a pandas dataframe using
to_pandas
df = filings.to_pandas()
Working with an individual Filing
Once you have retrieved Filings you can access individual filings using the bracket []
notation.
filings = get_filings(2021)
filing = filings[0]
Pay attention to the index value displayed for the filings. This is the value you will use to get the individual filing.
filing = filings[0]
Open a filing
You can open the filing in your browser using filing.open()
filing.open()
Open the Filing Homepage
You can open the filing homepage in the browser using filing.open_homepage()
filing.open()
Use DuckDB to query the filings
A conveient way to query the filings data is to use DuckDB. If you call the to_duckdb
function, you get an in-memory
DuckDB database instance, with the filings registered as a table called filings
.
Then you can work directy with the DuckDB database, and run SQL against the filings data.
In this example, we filter filings for S-1 form types.
db = filings.to_duckdb()
# a duckdb.DuckDBPyConnection
# Query the filings for S-1 filings and return a dataframe
db.execute("""
select * from filings where Form == 'S-1'
""").df()
Contributing
Contributions are welcome! We would love to hear your thoughts on how this library could be better at working with SEC Edgar.
Reporting Issues
We use GitHub issues to track public bugs. Report a bug by opening a new issue; it's that easy!
Making code changes
- Fork the repo and create your branch from master.
- If you've added code that should be tested, add tests.
- If you've changed APIs, update the documentation.
- Ensure the test suite passes.
- Make sure your code lints.
- Issue that pull request!
License
edgartools
is distributed under the terms of the MIT license.
Contact
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for edgartools-1.4.3-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2966558d6b220fb6d96e48dde2a4b1f2caaad9c42b8bcb8fa7e685f68b7c75d1 |
|
MD5 | 31314bc3063d024c7c529b4bc04a3ec8 |
|
BLAKE2b-256 | c0c6d3a078019cde7399a3f7ee6c2554469b25c88dad4e0a222fa47cbfe7ba67 |