README.MD

Documentation

Table of Contents

1. List of classes & functions
2. Identifying file types
3. Parsing Output data structure
4. How to get data

List of classes & functions

Package pysecxbrl.extract

Class pysecxbrl.extract.XBRLExtractor

Identifying file types

Parsing Output data structure

The current version will give you 2 dictionaries, one for the data and another for the context.

Data part:

{"id":{ # the id of the data object
    "tag":"...", # name of the tag, e.g. Revenue, CostOfSales ...
    "value":"...", # value of the tag, current version returns all values in string
    "prefix":"...", # namespace in the XML, e.g. us-gaap, dei, ...
    "contextRef":"..." # the reference ID to the context part
    "...":"..." # here you can have other attributes specific to the data object
    }
}

Context part:

In the current version, 2 types of context are considered:

1. unit of measure (USD, CAD, ...)
2. context: time + segment (time instant/stard & end date, which part of the business or other stakeholders)

unit

{"id":{ # the id of the context object
    "type":"unit", # the unit type
    "unit":"...", # value of the unit, e.g. can be USD
    "...":"..." # other attributes are possible
    }
}

context

{"id":{ # the id of the context object
    "type":"context", # the context type
    "instant":"2020-05-08", # it will be either a date or start+end
    "startDate":"...","endDate":"...",
    "segment":[ # list of one or more stakeholders
      "explicitMember":"...", # the "who"
      "dimension":"..." # can be very diverse
    ],
    "...":"..." # other attributes are possible
    }
}

How to get data

1. The entry point is here: https://www.sec.gov/edgar/searchedgar/accessing-edgar-data.htm
2. As is mentioned on the page, you can get daily filing lists here: https://www.sec.gov/Archives/edgar/daily-index/
3. Getting into the folders, you can see several types of index files, they are normally the same but organized in different ways, I personally prefer the "master" file
4. The "Master" file is a text file, and every line is like this: 1001115|GEOSPACE TECHNOLOGIES CORP|10-Q|20200508|edgar/data/1001115/0001564590-20-023322.txt
5. This text file is in fact not the XBRL file, but we can get its folder path: https://www.sec.gov/Archives/edgar/data/1001115
6. And in the inner folder, you can see the XBRL file which is 0001564590-20-023322-xbrl.zip