Skip to content

livMatS/dserver-dependency-graph-plugin

Repository files navigation

dserver Dependency Graph Plugin

dtool pypi tag test zenodo

Features

  • Use a dataset UUID to lookup all datasets within the same dependency graph
  • Build views on dataset dependency trees based on arbitrary connecting keys

Introduction

dtool is a command line tool for packaging data and metadata into a dataset. A dtool dataset manages data and metadata without the need for a central database.

However, if one has to manage more than a hundred datasets it can be helpful to have the datasets' metadata stored in a central server to enable one to quickly find datasets of interest.

The dservercore provides a web API for registering datasets' metadata and provides functionality to lookup, list and search for datasets.

This plugin enables the dserver to directly provide all datasets within a specific dependency graph.

Installation

Install the dserver dependency graph plugin

$ pip install dserver-dependency-graph-plugin

Setup and configuration

Configure plugin behavior

With

export DSERVER_ENABLE_DEPENDENCY_VIEW=True

the underlying database will offer a view on the default collection. This view offers an on-the-fly-generated collection of undirected per-dataset adjacency lists in order to facilitate searching dataset dependeny graphs in both directions. With

export DSERVER_FORCE_REBUILD_DEPENDENCY_VIEW=True

this view is reestablished at every query. This is required to apply changes to related options, such as the JSON-formatted list

export DSERVER_DEPENDENCY_KEYS='["readme.derived_from.uuid", "annotations.source_dataset_uuid"]'

which indicates at which keys to look for source dataset(s) by UUID. The example above illustrates the default. All keys are treated equivalently and nested fields are separated by the dot (.). The actual nesting hierarchy does not matter. This means, for example, both structures evaluate equivalently in the following

{'readme': {'derived_from': {'uuid':
    ['8ecd8e05-558a-48e2-b563-0c9ea273e71e',
     'faa44606-cb86-4877-b9ea-643a3777e021']}}}

{'readme': {'derived_from':
    [{'uuid': '8ecd8e05-558a-48e2-b563-0c9ea273e71e'},
     {'uuid': 'faa44606-cb86-4877-b9ea-643a3777e021'}]}}

Setting

export DSERVER_DYNAMIC_DEPENDENCY_KEYS=True

will allow the client side to request graphs spanned by arbitrary dependency keys (see below). The related options

export DSERVER_MONGO_DEPENDENCY_VIEW_PREFIX=dep
export DSERVER_MONGO_DEPENDENCY_VIEW_BOOKKEEPING=dep_views
export DSERVER_MONGO_DEPENDENCY_VIEW_CACHE_SIZE=10

control internal behavior. See source code and examples below.

Note that the above exports containing JSON syntax are formatted for usage in bash. Enclosing single quotes are not to be part of the actual variable value when environment variables are configured elsewhere.

dserver API

dserver makes use of the Authorized header to pass through the JSON web token for authorization. Below we create environment variables for the token and the header used in the curl commands

$ TOKEN=$(flask user token olssont)
$ HEADER="Authorization: Bearer $TOKEN"

Standard user usage

Looking up dependency graphs based on a dataset's UUID

A dataset can be derived from one or several source datasets, usually by machine-generated annotations attached via the Python API at dataset creation time, or manually by recording the UUIDs of parent datasets in some arbitrary fields within the README.yml. If configured appropriately, querying the server directly for all datasets within the same dependency graph by UUID is possible, i.e.

$ UUID=41a2e3e2-0c01-444f-bd7d-f9bb45512373
$ curl -H "$HEADER" http://localhost:5000/graph/lookup/$UUID

Looking up a dependency graph by UUID will result in unique per-UUID hits. As it is possible for a dataset to be registered in more than one base URI, the query will yield one arbitrary hit in such a case.

Looking up graphs spanned by arbitrary dependency keys

If DSERVER_DYNAMIC_DEPENDENCY_KEYS=True, then the client may ask the server to explore a graph spanned by dependency keys differing from the server-side defaults in DSERVER_DEPENDENCY_KEYS. This happens as above, but with via a POST request with a JSON-formatted list of desired dependency keys attached

$ curl -H "$HEADER" -H "Content-Type: application/json"  \
    -X POST -d  \
    '["annotations.source_dataset_uuid","readme.derived_from.uuid"]'
    http://localhost:5000/graph/lookup/$UUID

If a view for this particular set of keys does not exist yet, the server will generate and cache it on-the-fly. This can be observed in the mongo shell

$ mongo

> show dbs
admin       0.000GB
config      0.000GB
dtool_info  0.020GB
local       0.000GB

> use dtool_info
switched to db dtool_info

> show collections
datasets
dep:2020-10-05T01:22:39.581592
dep:2020-10-06T21:45:00.525410
dep:2020-10-06T21:45:28.495903
dep_views
dependencies
system.views

Here, all dep-prefixed collections are dependency views for distinct sets of keys. The bookkeeping collection``dep_views`` holds records of all dependency view - key set mappings together with the latest access

> db.dep_views.find()
{ "_id" : ObjectId("5f7a755faea9fcf239f68dba"), "name" : "dep:2020-10-05T01:22:39.581592", "keys" : [ "annotations.source_dataset_uuid", "readme.derived_from.uuid" ], "accessed_on" : ISODate("2020-10-07T12:24:32.724Z") }
{ "_id" : ObjectId("5f7ce55caea9fcf239f68dbb"), "name" : "dep:2020-10-06T21:45:00.525410", "keys" : [ "readme.derived_from.uuid" ], "accessed_on" : ISODate("2020-10-06T21:45:00.538Z") }
{ "_id" : ObjectId("5f7ce578aea9fcf239f68dbc"), "name" : "dep:2020-10-06T21:45:28.495903", "keys" : [ "bla" ], "accessed_on" : ISODate("2020-10-06T21:45:28.498Z") }

and querying with a specific set of keys for the first time

$ curl -H "$HEADER" -H "Content-Type: application/json"  \
    -X POST -d  \
    '["another.possibly_nested.dependency_key"]'  \
    http://localhost:5000/graph/lookup/$UUID

will result in an additional view named uniquely by the current UTC time:

> show collections
datasets
dep:2020-10-05T01:22:39.581592
dep:2020-10-06T21:45:00.525410
dep:2020-10-06T21:45:28.495903
dep:2020-10-07T17:03:58.831223
dep_views
dependencies
system.views

and an according entry within dep_views

> db.dep_views.find()
{ "_id" : ObjectId("5f7a755faea9fcf239f68dba"), "name" : "dep:2020-10-05T01:22:39.581592", "keys" : [ "annotations.source_dataset_uuid", "readme.derived_from.uuid" ], "accessed_on" : ISODate("2020-10-07T16:59:12.467Z") }
{ "_id" : ObjectId("5f7ce55caea9fcf239f68dbb"), "name" : "dep:2020-10-06T21:45:00.525410", "keys" : [ "readme.derived_from.uuid" ], "accessed_on" : ISODate("2020-10-06T21:45:00.538Z") }
{ "_id" : ObjectId("5f7ce578aea9fcf239f68dbc"), "name" : "dep:2020-10-06T21:45:28.495903", "keys" : [ "bla" ], "accessed_on" : ISODate("2020-10-06T21:45:28.498Z") }
{ "_id" : ObjectId("5f7df4feaea9fcf239f68dbd"), "name" : "dep:2020-10-07T17:03:58.831223", "keys" : [ "another.possibly_nested.dependency_key" ], "accessed_on" : ISODate("2020-10-07T17:03:58.833Z") }

If the total number of such cached views exceeds the allowed maximum value configured in DSERVER_MONGO_DEPENDENCY_VIEW_CACHE_SIZE, then the view accessed least recently is dropped.

About

dserver plugin for querying dataset dependency graphs

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages