changelog shortlog graph tags branches changeset file revisions annotate raw help

Mercurial > demo / examples/vegadat.lisp

revision 37: c6d0a37a046a
child 38: 8259376eee11
     1.1--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2+++ b/examples/vegadat.lisp	Sun Dec 24 19:24:39 2023 -0500
     1.3@@ -0,0 +1,38 @@
     1.4+;; from https://github.com/Lisp-Stat/plot/blob/master/src/vega/vega-datasets.lisp
     1.5+(defpackage :examples/vegadat
     1.6+  (:use :cl :std :net/fetch :dat)
     1.7+  (:export 
     1.8+   :*vega-datasets* :*vega-dataset-base-url*
     1.9+   :fetch-vega-dataset))
    1.10+
    1.11+(in-package :examples/vegadat)
    1.12+
    1.13+(defparameter *vega-dataset-base-url* "http://raw.githubusercontent.com/vega/vega-datasets/main/data/"
    1.14+  "Base URL for datasets included in Vega")
    1.15+
    1.16+(defparameter *vega-dataset-stash* "vega/")
    1.17+
    1.18+
    1.19+;; (gethash :airpots *vega-datasets*)
    1.20+(defvar *vega-datasets* (make-hash-table :size 66 :test #'equal)
    1.21+  "All Vega example data sets. k=symbol,v=url")
    1.22+
    1.23+(defun push-dataset (key)
    1.24+  "Push a dataset to *VEGA-DATASETS* by filename."
    1.25+  (let ((val (concatenate 'string *vega-dataset-base-url* key)))
    1.26+    (setf (gethash key *vega-datasets*) val)))
    1.27+
    1.28+;; 66 files total, mostly json and csv. 1 tsv file, 1 arrow file.
    1.29+(mapc #'push-dataset
    1.30+      '("airports.csv" "annual-precip.json" "anscombe.json" "barley.json" "budget.json" "budgets.json" "burtin.json" "cars.json" "countries.json" "crimea.json" "driving.json" "earthquakes.json" "flare-dependencies.json" "flare.json" "flights-10k.json" "flights-200k.json" "flights-20k.json" "flights-2k.json" "flights-5k.json" "football.json" "gapminder.json" "income.json" "jobs.json" "londonBoroughs.json" "londonCentroids.json" "londonTubeLines.json" "miserables.json" "monarchs.json" "movies.json" "normal-2d.json" "obesity.json" "ohlc.json" "penguins.json" "points.json" "political-contributions.json" "population.json" "udistrict.json" "unemployment-across-industries.json" "uniform-2d.json" "us-10m.json" "us-state-capitals.json" "volcano.json" "weather.json" "wheat.json" "world-110m.json" "airports.csv" "birdstrikes.csv" "co2-concentration.csv" "disasters.csv" "flights-3m.csv" "flights-airport.csv" "gapminder-health-income.csv" "github.csv" "iowa-electricity.csv" "la-riots.csv" "lookup_groups.csv" "lookup_people.csv" "population_engineers_hurricanes.csv" "seattle-weather-hourly-normals.csv" "seattle-weather.csv" "sp500-2000.csv" "sp500.csv" "stocks.csv" "us-employment.csv" "weather.csv" "windvectors.csv" "zipcodes.csv" "unemployment.tsv" "flights-200k.arrow"))
    1.31+
    1.32+(defun fetch-vega-datasets ()
    1.33+  (ensure-directories-exist *vega-dataset-stash*)
    1.34+  (maphash-keys
    1.35+   (lambda (x) (download (gethash x *vega-datasets*) 
    1.36+                         (merge-pathnames x *vega-dataset-stash*)))
    1.37+   *vega-datasets*))
    1.38+
    1.39+(defun purge-vega-datasets ()
    1.40+  (std:when-let ((stash (probe-file *vega-dataset-stash*)))
    1.41+    (sb-ext:delete-directory stash :recursive t)))