Clean downloads folder before creating input data zip.

Fixes #82

2 files changed, 20 insertions, 2 deletions

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 3008b1a..a9b0827 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -10,6 +10,7 @@
 
 * Switch command line flag processing to use the yargs library. Existing flags have changed: --no-validation and --filtered-zones have been renamed to --no_validation and --included_zones respectively. --included_zones now takes a list without quotes or commas.
 * Addition of new flags: --excluded_zones, --dist_dir, --downloads_dir, --skip_zip, --skip_shapefile. See --help and README.md for details.
+* Remove unneeded downloaded files from downloads directory before creating input data zipfile ([#82](https://github.com/evansiroky/timezone-boundary-builder/issues/82)). 
 
 ## 2020a
 
diff --git a/index.js b/index.js
index fa747b3..eef7a8f 100644
--- a/index.js
+++ b/index.js
@@ -1,5 +1,6 @@
 var exec = require('child_process').exec
 var fs = require('fs')
+var path = require('path')
 
 var area = require('@mapbox/geojson-area')
 var geojsonhint = require('@mapbox/geojsonhint')
@@ -12,7 +13,6 @@ var jsts = require('jsts')
 var rimraf = require('rimraf')
 var overpass = require('query-overpass')
 var yargs = require('yargs')
-var path = require('path')
 
 const ProgressStats = require('./progressStats')
 
@@ -723,7 +723,24 @@ const autoScript = {
     overallProgress.beginTask('Downloading osm boundaries')
     asynclib.eachSeries(Object.keys(osmBoundarySources), downloadOsmBoundary, cb)
   }],
-  zipInputData: ['makeDistDir', 'getOsmBoundaries', function (results, cb) {
+  cleanDownloadFolder: ['makeDistDir', 'getOsmBoundaries', function (results, cb) {
+    overallProgress.beginTask('cleanDownloadFolder')
+    const downloadedFilenames = Object.keys(osmBoundarySources).map(name => `${name}.json`)
+    fs.readdir('downloads', (err, files) => {
+      if (err) return cb(err)
+      asynclib.each(
+        files,
+        (file, fileCb) => {
+          if (downloadedFilenames.indexOf(file) === -1) {
+            return fs.unlink(path.join('downloads', file), fileCb)
+          }
+          fileCb()
+        },
+        cb
+      )
+    })
+  }],
+  zipInputData: ['cleanDownloadFolder', function (results, cb) {
     overallProgress.beginTask('Zipping up input data')
     exec('zip ' + distDir + '/input-data.zip ' + downloadsDir +
          '/* timezones.json osmBoundarySources.json expectedZoneOverlaps.json', cb)