riptano / cdm Goto Github PK

Context.download(url, cache=True)

connecting to Cassandra / DSE should really be one of the first things that happens

related to #36

graph.... groovy JONGRAPHLANG?

questionable.

if it's not present, error out with message

Need download() for testing

should be printable. easy to add elements of different types

schema.add(element)

elements should be fetchable by:

schema.find(Type)

or something similar

fail to install if running cassandra version doesn't meet the version requirements specified in the metadata

should be gremlin (host/graph) >

use --loglevel or whatever is the convention

it should be trivial to provide feedback to the user of what's going on. Context.feedback() will be a standardized means of displaying feedback to the console or web.

Installing dataset movielens-small==master
Traceback (most recent call last):
File "/Users/jhaddad/.virtualenvs/cdm/bin/cdm", line 9, in
load_entry_point('Cassandra-Data-Manager==0.1', 'console_scripts', 'cdm')()
File "/Users/jhaddad/dev/cdm/cdm/cli.py", line 53, in main
install_search=arguments['--search'])
File "/Users/jhaddad/dev/cdm/cdm/util.py", line 71, in install
repo = download_dataset(dataset, y[dataset]['url'])
KeyError: 'movielens-small'

Code examples:

Create vertex & edge. LABEL is optional.

CREATE vertex movie
CREATE VERTEX LABEL person
CREATE VERTEX LABEL movie
CREATE edge rated

Creating vertex and edge properties. types should work with all CQL primitive types:

CREATE PROPERTY name text
CREATE PROPERTY person_id INTEGER

Indexes:

CREATE INDEX user_id_idx ON movie(user_id) MATERIALIZED
CREATE INDEX search on movie(title) FULLTEXT
CREATE INDEX movie_title_idx ON movie(title) SEARCH;

Groovy code:

reviewer.buildEdgeIndex('ratedByStars', rated).direction(OUT).byPropertyKey('stars').add()

graph.schema().vertexLabel('reviewer').buildEdgeIndex('ratedByStars', rated).direction(OUT). byPropertyKey('stars').add()

CREATE [direction] INDEX [name] ON EDGE [label]([vertex_label.property])

CREATE OUT index ratedbyStars on edge rated(reviewer.stars)

For now, just host and port would be sufficient. But for the future, it would be extremely useful to not setup both cqlsh and cdm separately :)

create vertex person, place, thing
create edge knows, likes, despises

build webapp to control cdm from the browser

if post_install.py is present, it'll attempt to run the main() inside, passing a Context object, which has db connection (set to correct keyspace). main() should load any data into the DB.

Edit readme with details on post_install.py before closing this issue

show warning if it doesn't exist

needs to run through the full schema setup

work with the normal cassandra-driver, but if dse module is available it should use that as a priority

I should be able to tag datasets with things like materialized views, sasi, graph, etc. list of tags should be standardized.

should expose a context that has everything someone would need available to generate a dataset

utilities to download temp files (and cache)
cassandra cluster and session