ckanapi-exporter

Export dataset metadata from CKAN to Excel


Keywords
ckan, csv, export, excel, api
License
Other
Install
pip install ckanapi-exporter==0.1.0

Documentation

Build Status Coverage Status Latest Version Supported Python versions Development Status License

ckanapi-exporter

An API client (usable as a command-line script or as a Python library) for exporting dataset metadata from CKAN sites to Excel-compatible CSV files.

Installation

To install run:

pip install ckanapi-exporter

Usage

ckanapi-exporter --url 'https://demo.ckan.org' \
    --column "Title" --pattern '^title$' > output.csv

This searches each dataset on demo.ckan.org for fields matching the regular expression ^title$ (the --pattern argument) and puts the values into a column called "Title" in the CSV file (the --column argument). It'll create an output.csv file something like this:

Title
Senior Salaries Information
Demo Data for Open Data in 1 Day - Spending Over £500
UK Cat Burglaries
...

You can add as many columns as you want: just add a --column and a --pattern argument for each column. The title of the column in the CSV file can be anything you want - it doesn't have to match the name of the field in CKAN. Let's add a second column titled "Rights" that contains the license_title fields from the datasets:

ckanapi-exporter --url 'https://demo.ckan.org' \
    --column "Title" --pattern '^title$' \
    --column "Rights" --pattern '^license_title$' > output.csv
Title Rights
Senior Salaries Information Creative Commons Attribution
Demo Data for Open Data in 1 Day - Spending Over £500 Creative Commons CCZero
UK Cat Burglaries UK Open Government Licence (OGL)
... ...

API Parameters

The ckanapi-exporter calls the package_search API action and you can pass in related query parameters by using the --params argument and passing in a string formated as a dictionary. Each key: value pair represents a query passed to the API call.

For example if you wanted to only export datasets between a date range you can pass in the fq (filtered query) parameter and use metadata_created to filter the results.

ckanapi-exporter --url 'https://demo.ckan.org' \
    --params "{'fq':'metadata_created:[2017-01-01T00:00:00Z TO 2017-01-31T23:59:99.999Z]'}" \
    --column "Title" --pattern '^title$' \
    --column "Rights" --pattern '^license_title$' > output.csv

Transformations

You can apply certain transformations to the values from the datasets. For example, let's add a third column with the first 50 characters of each dataset's description (the notes field in the CKAN API):

ckanapi-exporter --url 'https://demo.ckan.org' \
    --column "Title" --pattern '^title$' \
    --column "Rights" --pattern '^license_title$' \
    --column "Description" --pattern '^notes$' --max-length 50 > output.csv
Title Rights Description
Senior Salaries Information Creative Commons Attribution Demo information about senior salaries from 11/04/
Demo Data for Open Data in 1 Day - Spending Over £500 Creative Commons CCZero Data on spending over £500 generated for Open Data
UK Cat Burglaries UK Open Government Licence (OGL) A record of cat burgalries, listing the cat names,
... ... ...

Exporting Resource Fields

Let's add a column containing the formats of each datasets' resources:

ckanapi-exporter --url 'https://demo.ckan.org' \
    --column "Title" --pattern '^title$' \
    --column "Rights" --pattern '^license_title$' \
    --column "Description" --pattern '^notes$' --max-length 50 \
    --column Formats --pattern '^resources$' '^format$' > output.csv

This time the pattern has two arguments: --pattern '^resources$' '^format$'. This means find the "resources" field of each dataset and then find the "format" field of each resource. When a dataset has more than one resource the formats will be combined into a quoted, comma-separated list in a single table cell. It'll create a CSV file something like this:

Title Rights Description Formats
Senior Salaries Information Creative Commons Attribution Demo information about senior salaries from 11/04/ XLSX, CSV
Demo Data for Open Data in 1 Day - Spending Over £500 Creative Commons CCZero Data on spending over £500 generated for Open Data CSV, CSV, CSV, CSV
UK Cat Burglaries UK Open Government Licence (OGL) A record of cat burgalries, listing the cat names, JPEG, CSV, CSV
... ... ... ...

CSV is repeated a lot because lots of the datasets have multiple CSV resources. You can add the --deduplicate option to the column to remove the duplication:

ckanapi-exporter --url 'https://demo.ckan.org' \
    --column "Title" --pattern '^title$' \
    --column "Rights" --pattern '^license_title$' \
    --column "Description" --pattern '^notes$' --max-length 50 \
    --column Formats --pattern '^resources$' '^format$' --deduplicate \
    > output.csv
Title Rights Description Formats
Senior Salaries Information Creative Commons Attribution Demo information about senior salaries from 11/04/ XLSX, CSV
Demo Data for Open Data in 1 Day - Spending Over £500 Creative Commons CCZero Data on spending over £500 generated for Open Data CSV
UK Cat Burglaries UK Open Government Licence (OGL) A record of cat burgalries, listing the cat names, JPEG, CSV
... ... ... ...

Exporting Dataset Extras

Let's add a column with the values of the "Next Update" extra from each dataset. Dataset publishers have been inconsistent with naming this column, it's sometimes "Next Update" and sometimes "next update", "Next update day", "Next Update Time" etc. We'll use a regular expression that matches all of these possible names and combine them into a single "Next Update" column:

ckanapi-exporter --url 'https://demo.ckan.org' \
    --column "Title" --pattern '^title$' \
    --column "Rights" --pattern '^license_title$' \
    --column "Description" --pattern '^notes$' --max-length 50 \
    --column Formats --pattern '^resources$' '^format$' --deduplicate \
    --column "Next Update" --pattern '^extras$' '^next update.*' --unique \
    > output.csv

The two-part pattern '^extras$' '^next update.*' means to look in the "extras" field of each dataset for extras whose name matches ^next update.*. We're expecting each dataset to have only one matching extra so we add the --unique argument which will crash if a dataset has more than one extra matching the pattern.

By default patterns are matched case-insensitively and whitespace is stripped from field names before matching. To match case-sensitively and without stripping whitespace add --case-sensitive --strip false to the column.

We can also find multiple extras and combine them into a single column. For example, let's say our datasets have a "contributor" extra (sometimes spelled "contributor", sometimes "Contributor"). Some datasets have multiple extras named "Contributor 1", "Contributor 2" etc. We can find all of these contributor extras and combine them into a single quoted, comma-separated list with a pattern like this:

--column Contributors --pattern '^extras$' '^contributor.*'

Using a columns.json File

You can specify your columns in a columns.json file instead of on the command line. Here's an example of the format:

{
  "Data Owner": {
      "pattern": "^author$",
      "unique": true,
      "case_sensitive": true
  },
  "Delivery Unit": {
      "pattern": ["^extras$", "^Delivery Unit$"],
      "unique": true
  },
  "Contributor": {
      "pattern": ["^extras$", "^Contributor.*"]
  },
  "Description": {
      "pattern": "^notes$",
      "unique": true,
      "case_sensitive": true,
      "max_length": 255
  },
  "Format": {
      "pattern": ["^resources$", "^format$"],
      "case_sensitive": true,
      "deduplicate": true
  }
}

Then tell ckanapi-exporter to read the column options from this file instead of giving them on the command line:

ckanapi-exporter --url 'https://demo.ckan.org' --columns columns.json > output.csv

For a working example columns.json file that you can use against demo.ckan.org, see test_columns.json.

ckanapi-exporter is a thin wrapper around losser, hooking it up to the CKAN API. For more documentation of the filtering and transforming options run ckanapi-exporter --help or read losser's docs.

Using as a Python Library

You can also import ckanapi-exporter in Python and use it from your CKAN API client or plugin:

import ckanapi_exporter.exporter as exporter
csv_string = exporter.export('https://demo.ckan.org', 'columns.json')

Returns a UTF8-encoded string.

The second argument can be either the filename of the columns.json file as a string, or a list of dictionaries (equivalent to the contents of columns.json file after loading the JSON).

Development

To install for development, create and activate a Python virtual environment then do:

git clone https://github.com/ckan/ckanapi-exporter.git
cd ckanapi-exporter
python setup.py develop
pip install -r dev-requirements.txt

To run the tests do:

nosetests