energy-dashboard-client 0.4

Energy Dashboard Command Line Interface (CLI)

Energy Dashboard Command Line Interface (edc)

Command Line Interface for the Energy Dashboard.

!!!PRE-ALPHA!!!

While this is the master branch, this project is not released yet. Stand by...

All examples commands, install, etc. assume a linux (ubuntu) installation and use the apt package manager, etc.

Prerequisites

Install basic deps

sudo apt install parallel
sudo apt install build-essential
sudo apt install git

Install git-lfs (git large file store)

git-lfs is used for storing the database files, which are basically binary blobs that are updated periodically. Rather than store all the db blob revisions in the git repository, which would bloat it considerably, the db blobs are offloaded to git-lfs.

For installation instructions, go here:

• https://git-lfs.github.com/

Example:

curl -s https://packagecloud.io/install/repositories/github/git-lfs/script.deb.sh | sudo bash

Install conda/anaconda

You don't strictly need anaconda for this toolchain to work. If you prefer mucking with python virtualenv directly, then go for it. I find that anaconda works really well with other parts of this toolchain, namely Jupyter Notebooks. All the examples and documentation will assume you are using anaconda.

Example, see the website for current instructions:

• https://www.anaconda.com/distribution/#download-section

Example:

wget https://repo.anaconda.com/archive/Anaconda3-2019.07-Linux-x86_64.sh
chmod +x Anaconda3-2019.07-Linux-x86_64.sh 
./Anaconda3-2019.07-Linux-x86_64.sh 

Installation

This webpage has a great tutorial on how to use conda. It's what I use when I forget the commands and concepts:

• https://geohackweek.github.io/Introductory/01-conda-tutorial/

First, create a conda environment, it can be named anything, I'll call this edc-cli:

conda update conda
conda create -n edc-cli python=3 numpy jupyter pandas
conda activate edc-cli

Then install the energy-dashboard-client:

pip install -U energy-dashboard-client

Setup

The energy-dashboard-client has two commands to get you up and running with an energy-dashboard:

• clone : this will literally use git to clone the energy-dashboard repo to your local machine
• update : this will pull down all the submodules to your local machine

Note: if you only want a subset of the submodules installed on your local machine, then you can use the git submodule deinit data/[name-of-submodule-to-remove].

As always, let me know if you need better tooling around this or any other aspect of this project.

mkdir foo
cd foo
edc clone
cd energy-dashboard
edc update

At this point you should have a working environment:

Verify that you have files:

$ tree -L 1
.
├── data
├── docs
├── LICENSE
├── notebooks
├── README.md
└── run.sh

Verify that edc works:

`$ edc --help
Usage: edc [OPTIONS] COMMAND [ARGS]...

  Command Line Interface for the Energy Dashboard. This tooling  collects
  information from a number of data feeds, imports that data,  transforms
  it, and inserts it into a database.

Options:
  --ed-dir TEXT                   Energy Dashboard directory (defaults to cwd)
  --log-level [CRITICAL|ERROR|WARNING|INFO|DEBUG]
  --help                          Show this message and exit.

Commands:
  clone    Clone energy-dashboard locally
  feed     Manage individual 'feed' (singular).
  feeds    Manage the full set of data 'feeds' (plural).
  license  Show the license (GPL v3).
  update   Update the submodules

Verify that you can list out the data feeds:

$ edc feeds list | head
data-oasis-atl-ruc-zone-map
data-oasis-cbd-nodal-grp-cnstr-prc
data-oasis-cmmt-rmr-dam
data-oasis-atl-sp-tie
data-oasis-prc-mpm-cnstr-cmp-dam
data-oasis-trns-curr-usage-all-all
data-oasis-ene-baa-mkt-events-rtd-all
data-oasis-ene-eim-transfer-limit-all-all
data-oasis-as-results-dam
data-oasis-ene-wind-solar-summary

Using find we can verify that we don't have any data files such as .zip, .xml, or .sql in the tree, but that we do have the state files:

$ find data/ | grep state | head
data/data-oasis-atl-ruc-zone-map/sql/state.txt
data/data-oasis-atl-ruc-zone-map/xml/state.txt
data/data-oasis-atl-ruc-zone-map/zip/state.txt
data/data-oasis-cbd-nodal-grp-cnstr-prc/sql/state.txt
data/data-oasis-cbd-nodal-grp-cnstr-prc/xml/state.txt
data/data-oasis-cbd-nodal-grp-cnstr-prc/zip/state.txt
data/data-oasis-cmmt-rmr-dam/sql/state.txt
data/data-oasis-cmmt-rmr-dam/xml/state.txt
data/data-oasis-cmmt-rmr-dam/zip/state.txt
data/data-oasis-atl-sp-tie/sql/state.txt

Now verify what databases you have downloaded...

$ find data/ | grep "\.db$" | head
data/data-oasis-atl-gen-cap-lst/db/data-oasis-atl-gen-cap-lst_00.db
data/data-oasis-sld-adv-fcst-rtd/db/data-oasis-sld-adv-fcst-rtd_01.db
data/data-oasis-sld-adv-fcst-rtd/db/data-oasis-sld-adv-fcst-rtd_00.db
data/data-oasis-sld-sf-eval-dmd-fcst/db/data-oasis-sld-sf-eval-dmd-fcst_00.db
data/data-oasis-sld-sf-eval-dmd-fcst/db/data-oasis-sld-sf-eval-dmd-fcst_03.db
data/data-oasis-sld-sf-eval-dmd-fcst/db/data-oasis-sld-sf-eval-dmd-fcst_01.db
data/data-oasis-sld-sf-eval-dmd-fcst/db/data-oasis-sld-sf-eval-dmd-fcst_05.db
data/data-oasis-sld-sf-eval-dmd-fcst/db/data-oasis-sld-sf-eval-dmd-fcst_04.db
data/data-oasis-sld-sf-eval-dmd-fcst/db/data-oasis-sld-sf-eval-dmd-fcst_02.db
data/data-oasis-ene-flex-ramp-dc-rtd-all/db/data-oasis-ene-flex-ramp-dc-rtd-all_00.db

I'll go over this in more detail below, but the reason there are multiple database files for a given data feed is because the feed has multiple formats (argh!) and I have not yet sorted out how to deal with that. More on this later.

Use Cases

Create Jupyter Notebook

TODO

This is what most of the users of this project want to do.

Process Data Feeds

At a high level, a data feed is simply a url and some instructions for processing it. The url is stored in the manifest.json, and the processing instructions are stored in the ./src directory. The ./src directory contains python files that handle downloading, parsing, constructing sql insert statements, and inserting the data into a sqlite3 database. See the section on Add New Data Feed for more details on the construction of a data feed.

Data feeds are processed in stages, and you can think of this as a vertical to horizontal processing.

Horizontally, the process is very simple, we move from downloading a resource through the stages until we insert the records into a database. The DATABASE is the final product.

DOWNLOAD -> EXTRACT -> PARSE -> SQL -> INSERT -> *DATABASE*

Vertically, the process is also very simple. Each stage processes all the artifacts from the previous stage. However, we need this to be robust and re-startable. No idea if that's a word, but it's a real thing. Machines crash, or you may need to stop for some reason. I've also had the case where I had an error in the logic in a given stage, and needed to start over from scratch. How do we do this here? Easy. A state file:

DOWNLOAD        -> EXTRACT     -> PARSE SQL     -> INSERT       -> *DATABASE*
./zip/          ./xml/          ./sql/          ./db/
    state.txt       state.txt       state.txt       state.txt

Each state.txt contains a list of artifacts that have already been processed. Originally I called these: ./zip/downloaded.txt, ./xml/unzipped.txt, etc. But after working with this for a few days, it is easier to just cat out [dir]/state.txt rather than remembering what each state file is named.

Each stage in the pipeline looks at the artifacts in the previous stage and compares that list with the list of previously processed artifacts in it's state.txt file, and then gives the delta of new files to the processing code.

So, to restart a given stage, you just delete the stage directory. This deletes all the generated artifacts and the state file. Voila. You are ready to start over.

Note: if you delete a stage, you may want to delete the subsequent stages, too.

Here's the command that does this for you:

edc feed [feed name] reset [stage]

And to process a given stage:

edc feed [feed name] proc [stage]

Here's the scenario. I've been writing the code for this project on my laptop. But it does not have the horsepower to crunch all this data into the various sqlite databases in a reasonable amount of time. So I'm firing up a desktop machine to perform the heavy lifting. Here's what that process looks like. This is the same process any researcher that wanted to replicate my work would want to do.

Clone and update

mkdir foo
edc clone
cd energy-dashboard
edc update

At this point, we have the energy-dashboard project, but we don't want to re-download all the previously downloaded files from their original source. In the case of CAISO OASIS, that would simply take too long (I've calculated the upper bound as 152 days, though in reality it took about 3 weeks to download the resources here). Instead, we can pull these previously downloaded artifacts from one of the public S3 buckets that I've mirrored them on...

Example:

#edc feed [feed-name] s3restore
edc feed data-oasis-atl-ruc-zone-map s3restore

To grab the artifacts from the entire set of feeds:

edc feeds list | xargs -L 1 -I {} edc feed {} s3restore

Add New Data Feed

TODO

Show Help

edc

Usage: edc [OPTIONS] COMMAND [ARGS]...

  Command Line Interface for the Energy Dashboard. This tooling  collects
  information from a number of data feeds, imports that data,  transforms
  it, and inserts it into a database.

Options:
  --config-dir TEXT     Config file directory
  --debug / --no-debug  Enable debug logging
  --help                Show this message and exit.

Commands:
  config   Manage config file.
  feed     Manage individual 'feed' (singular).
  feeds    Manage the full set of data 'feeds' (plural).
  license  Show the license (GPL v3).

config

Usage: edc config [OPTIONS] COMMAND [ARGS]...

  Manage config file.

Options:
  --help  Show this message and exit.

Commands:
  show    Show the config
  update  Update config

feed

Usage: edc feed [OPTIONS] COMMAND [ARGS]...

  Manage individual 'feed' (singular).

Options:
  --help  Show this message and exit.

Commands:
  archive    Archive feed to tar.gz
  create     Create new feed
  download   Download from source url
  invoke     Invoke a shell command in the feed directory
  proc       Process a feed through the stages
  reset      Reset feed to reprocess stage
  restore    Restore feed from tar.gz
  s3archive  Archive feed to S3 bucket
  s3restore  Restore feed zip files from from S3 bucket
  status     Show feed status

feeds

Usage: edc feeds [OPTIONS] COMMAND [ARGS]...

  Manage the full set of data 'feeds' (plural).

Options:
  --help  Show this message and exit.

Commands:
  list    List feeds
  search  Search feeds (NYI)

license

    edc : Energy Dashboard Command Line Interface
    Copyright (C) 2019  Todd Greenwood-Geer (Enviro Software Solutions, LLC)

    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation, either version 3 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <https://www.gnu.org/licenses/>.

##Usage

Examples

edc feed invoke data-oasis-atl-lap-all "git st"
edc feed invoke data-oasis-atl-lap-all "ls"
edc feed invoke data-oasis-atl-lap-all "cat manifest.json"
edc feed invoke data-oasis-atl-lap-all "head manifest.json"
edc feeds list
edc feeds list | grep atl
edc feeds list | grep atl | edc feed invoke "head manifest.json"
edc feeds list | grep atl | edc feed invoke "head manifest.json" -
edc feeds list | grep atl | xargs -L 1 -I {} edc feed invoke {} "head manifest.json"
edc feeds list | grep atl | xargs -L 1 -I {} edc feed invoke {} "jq . < manifest.json"
edc feeds list | grep atl | xargs -L 1 -I {} edc feed invoke {} "jq .url < manifest.json"
edc feeds list | grep mileage | xargs -L 1 -I {} edc feed invoke {} "echo {}; sqlite3 db/{}.db 'select count(*) from oasis'"
edc feeds list | grep atl | xargs -L 1 -I {} edc feed invoke {} "jq .url < manifest.json"
edc feeds list| xargs -L 1 -I {} edc feed invoke {} "echo {}; sqlite3 db/{}.db 'select count(*) from oasis'"
edc feeds list | grep atl | xargs -L 1 -I {} edc feed status {}
edc feeds list | grep atl | xargs -L 1 -I {} edc feed status --header {}
edc feeds list | grep atl | xargs -L 1 -I {} edc feed status --header {}
edc feeds list | grep mileage | xargs -L 1 -I {} edc feed status --header {}
edc feeds list | xargs -L 1 -I {} edc feed invoke {} "./src/10_down.py"
edc feed archive data-oasis-as-mileage-calc-all
edc feed archive data-oasis-as-mileage-calc-all | xargs -L 1 -I {} tar -tvf {}
edc feed reset data-oasis-as-mileage-calc-all --stage xml --stage db
edc feed s3restore data-oasis-as-mileage-calc-all --outdir=temp --service=wasabi
edc feed s3archive data-oasis-as-mileage-calc-all

Onboarding

Some quick notes on how I onboarded 'data-oasis-as-mileage-calc-all':

edc feed proc data-oasis-as-mileage-calc-all
edc feed s3archive data-oasis-as-mileage-calc-all --service wasabi
edc feed s3archive data-oasis-as-mileage-calc-all --service digitalocean
edc feed status data-oasis-as-mileage-calc-all --header
edc feed invoke data-oasis-as-mileage-calc-all "git st"
edc feed invoke data-oasis-as-mileage-calc-all "git log"
edc feed invoke data-oasis-as-mileage-calc-all "git show HEAD"

Author

Todd Greenwood-Geer (Enviro Software Solutions, LLC)

Notes

This project uses submodules, and this page has been useful: https://github.blog/2016-02-01-working-with-submodules/