Skip to content

ruarxive/apibackuper

Repository files navigation

title
apibackuper \-- a command-line tool to archive/backup API calls

apibackuper is a command line tool to archive/backup API calls. It's goal to download all data behind REST API and to archive it to local storage. This tool designed to backup API data, so simple as possible.

::: contents :::

::: section-numbering :::

History

This tool was developed optimize backup/archival procedures for Russian government information from E-Budget portal budget.gov.ru and some other government IT systems too. Examples of tool usage could be found in "examples" directory

Main features

  • Any GET/POST iterative API supported
  • Allows to estimate time required to backup API
  • Stores data inside ZIP container
  • Supports export of backup data as JSON lines file
  • Documentation
  • Test coverage

Installation

Linux

Most Linux distributions provide a package that can be installed using the system package manager, for example:

# Debian, Ubuntu, etc.
$ apt install apibackuper
# Fedora
$ dnf install apibackuper
# CentOS, RHEL, ...
$ yum install apibackuper
# Arch Linux
$ pacman -S apibackuper

Windows, etc.

A universal installation method (that works on Windows, Mac OS X, Linux, …, and always provides the latest version) is to use pip:

# Make sure we have an up-to-date version of pip and setuptools:
$ pip install --upgrade pip setuptools

$ pip install --upgrade apibackuper

(If pip installation fails for some reason, you can try easy_install apibackuper as a fallback.)

Python version

Python version 3.6 or greater is required.

Quickstart

This example is about backup of Russian certificate authorities. List of them published at e-trust.gosuslugi.ru and available via undocumented API.

$ apibackuper create etrust
$ cd etrust

Edit apibackuper.cfg as:

[settings]
initialized = True
name = etrust

[project]
description = E-Trust UC list
url = https://e-trust.gosuslugi.ru/app/scc/portal/api/v1/portal/ca/list
http_mode = POST
work_modes = full,incremental,update
iterate_by = page

[params]
page_size_param = recordsOnPage
page_size_limit = 100
page_number_param = page

[data]
total_number_key = total
data_key = data
item_key = РеестровыйНомер
change_key = СтатусАккредитации.ДействуетС

[storage]
storage_type = zip

Add file params.json with parameters used with POST requests

{"page":1,"orderBy":"id","ascending":false,"recordsOnPage":100,"searchString":null,"cities":null,"software":null,"cryptToolClasses":null,"statuses":null}

Execute command "estimate" to see how long data will be collected and how much space needed

$ apibackuper estimate full

Output:

Total records: 502
Records per request: 100
Total requests: 6
Average record size 32277.96 bytes
Estimated size (json lines) 16.20 MB
Avg request time, seconds 66.9260
Estimated all requests time, seconds 402.8947

Execute command "run" to collect the data. Result stored in "storage.zip"

$ apibackuper run full

Exports data from storage and saves as jsonl file called "etrust.jsonl"

$ apibackuper export jsonl etrust.jsonl

Config options

Example config file

[settings]
initialized = True
name = <name>
splitter = .

[project]
description = <description>
url = <url>
http_mode = <GET or POST>
work_modes = <combination of full,incremental,update>
iterate_by = <page or skip>

[params]
page_size_param = <page size param>
page_size_limit = <page size limit>
page_number_param = <page number>
count_skip_param = <key to iterate in skip mode>


[data]
total_number_key = <total number key>
data_key = <data key>
item_key = <item key>
change_key = <change key>

[follow]
follow_mode = <type of follow mode>
follow_pattern = <url prefix to follow links>
follow_data_key = <follow data item key>
follow_param = <follow param>
follow_item_key = <follow item key>

[files]
fetch_mode = <file fetch mode>
root_url = <file root url>
keys = <keys with file data>
storage_mode = <file storage mode>


[storage]
storage_type = zip
compression = True

settings

  • name - short name of the project
  • splitter - value of field splitter. Needed for rare cases when '.' is part of field name. For example for OData requests and '@odata.count' field

project

  • description - text that explains what for is this project
  • url - API endpoint url
  • http_mode - one of HTTP modes: GET or POST
  • work_modes - type of operations: full - archive everything, incremental - add new records only, update - collect changed data only
  • iterate_by - type of iteration of records. By 'page' - default, page by page or by 'skip' if skip value provided

params

  • page_size_param - parameter with page size
  • page_size_limit - limit of records provided by API
  • page_number_param = parameter with page number
  • count_skip_param - parameter for 'skip' type of iteration

data

  • total_number_key - key in data with total number of records
  • data_key - key in data with list of records
  • item_key - key in data with unique identifier of the record. Could be group of keys separated with comma
  • change_key - key in data that indicates that record changed. Could be group of keys separated with comma

follow

  • follow_mode - mode to follow objects. Could be 'url' or 'item'. If mode is 'url' than follow_pattern not used
  • follow_pattern - url pattern / url prefix for followed objects. Only for mode 'item''
  • follow_data_key - if object/objects are inside array, key of this array
  • follow_param - parameter used in 'item' mode
  • follow_item_key - item key

files

  • fetch_mode - file fetch mode. Could be 'prefix' or 'id'. Prefix
  • root_url - root url / prefix for files
  • keys - list of keys with urls/file id's to search for files to save
  • storage_mode - a way how files stored in storage/files.zip. By default 'filepath' and files storaged same way as they presented in url

storage

  • storage_type - type of local storage. 'zip' is local zip file is default one
  • compression - if True than compressed ZIP file used, less space used, more CPU time processing data

Usage

Synopsis:

$ apibackuper [flags] [command] inputfile

See also apibackuper --help.

Examples

Create project "budgettofk":

$ apibackuper create budgettofk

Estimate execution time for 'budgettofk' project. Should be called in project dir or project dir provided via -p parameter:

$ apibackuper estimate full -p budgettofk

Output

Total records: 12282
Records per request: 500
Total requests: 25
Average record size 1293.60 bytes
Estimated size (json lines) 15.89 MB
Avg request time, seconds 1.8015
Estimated all requests time, seconds 46.0536

Run project. Should be called in project dir or project dir provided via -p parameter

$ apibackuper run full

Export data from project. Should be called in project dir or project dir provided via -p parameter

$ apibackuper export jsonl hhemployers.jsonl -p hhemployers

Follows each object of downloaded data and does requests for each objects .. code-block:: bash

$ apibackuper follow continue

Downloads all files associated with API objects .. code-block:: bash

$ apibackuper getfiles

Advanced

TBD