distil-ingest

Dependencies

Requires the Go programming language binaries with the GOPATH environment variable specified and $GOPATH/bin in your PATH.

Installation

go get github.com/uncharted-distil/distil-ingest

Development

Clone the repository:

mkdir $GOPATH/src/github.com/unchartedsoftware
cd $GOPATH/src/github.com/unchartedsoftware
git clone [email protected]:uncharted-distil/distil-ingest.git

Install dependencies:

cd distil-ingest
make install

Build executable:

make build

Usage

The repository contains CLIs used to parse, and ingest 3M OpenML datasets (those with a name beginning with o_) into elasticsearch.

Merging training and target datasets:

• Download D3M datasets of interest from https://datadrivendiscovery.org/data and unzip.
• Update and ensure the arguments in ./merge_all.share correct
• Run ./merge_all.sh

Classifying merged datasets:

• Update and ensure the arguments in ./classify_all.share correct
• Run ./classify_all.sh

Ingesting merged and classified datasets:

• Update and ensure the arguments in ./ingest_all.share correct
• Run ./ingest_all.sh

Common Issues:

"EOF"

• The Elasticsearch instance does not have http.compression enabled.
• The mappings json argument is invalid, most likely missing a closing bracket

"No Elasticsearch node available"

• You are accessing an Elasticsearch instance that requires a VPN and it is not on.
• The Elasticsearch instance is temporarily down.

"dep: command not found":

• Cause: $GOPATH/bin has not been added to your $PATH.
• Solution: Add export PATH=$PATH:$GOPATH/bin to your .bash_profile or .bashrc.