gtfs_df 0.4.0 → 0.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 127f50187dd26c4824abd94a6a44f64242715bf9092072276cdf6c811e0c28a6
-  data.tar.gz: d5a6bd9b25830d0574c6c315712d1ec4e0b14959d25f6a1cf12f2e13eeb738a2
+  metadata.gz: 205ec058b41c5bd1d2b01ff3950d4cd2ebb20f304d02d5bcd3dd0c447b4e0a6e
+  data.tar.gz: 54cca637de421c26d2144df100f8430f5dddb639b3a37d1160dfbd9704630e33
 SHA512:
-  metadata.gz: 9312f173960b069a31bb2bb37368bd65ad5345d527d5a29818a7485c731bcb6f6c5fb74639403a3bcbd5d5704ae2bbf629ed5cdfe3f5f629f36ea719663356da
-  data.tar.gz: 19563c256e1cbe52a34eef47c31f6b50806a0cb07b773bacb015ad40730c6d804941aef1c936aaeb6c17ea49719ab7db64d1144eeb05c7c4bd1d83ebcd87e12f
+  metadata.gz: ef853b504ee701e77911259352d0057f6e327db8022370a1fa0dbaa597bfadcac7ac1c8fbe746c04ec4669d86384279a259ac35d4126f1483286937812cb7cce
+  data.tar.gz: 728d684dd02b653cc779aa5e8152766f5dccafcb097e8fca5810b90f994c442645119f32efb62c84d8fbf1645109f579a3c74f6eb88c9e21dd03faa72193cdd0

data/CHANGELOG.md

@@ -1,3 +1,12 @@
+## [0.4.1] - 2025-12-05
+
+### Added
+
+- handle extra whitespace in csvs
+
+### Maintenance
+
+- remove unused initializer format
 ## [0.4.0] - 2025-12-04
 
 ### Added

data/lib/gtfs_df/base_gtfs_table.rb

@@ -10,20 +10,21 @@ module GtfsDf
         if input.is_a?(Polars::DataFrame)
           input
         elsif input.is_a?(String)
-          # We need to account for extra columns due to: https://github.com/ankane/ruby-polars/issues/125
-          all_columns = Polars.scan_csv(input).columns
-          default_schema = all_columns.map { |c| [c, Polars::String] }.to_h
-          dtypes = default_schema.merge(self.class::SCHEMA)
-          Polars.read_csv(input, null_values: [""], dtypes:)
-        elsif input.is_a?(Array)
-          head, *body = input
-          df_input = body.each_with_object({}) do |row, acc|
-            head.each_with_index do |column, index|
-              acc[column] ||= []
-              acc[column] << row[index]
-            end
-          end
-          Polars::DataFrame.new(df_input, schema_overrides: self.class::SCHEMA, strict: false)
+          # TODO: use `infer_schema: false` instead of `infer_schema_length` after polars release:
+          # https://github.com/ankane/ruby-polars/blob/master/CHANGELOG.md#100-unreleased
+          df = Polars.read_csv(input, infer_schema_length: 0)
+          dtypes = self.class::SCHEMA.slice(*df.columns)
+
+          df
+            .with_columns(dtypes.keys.map do |col|
+              stripped = Polars.col(col).str.strip
+              Polars.when(stripped.str.len_chars.gt(0))
+                .then(stripped)
+                .otherwise(Polars.lit(nil))
+            end)
+            .with_columns(dtypes.map do |name, type|
+                            Polars.col(name).cast(type)
+                          end)
         else
           throw GtfsDf::Error, "Unrecognized input"
         end

data/lib/gtfs_df/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module GtfsDf
-  VERSION = "0.4.0"
+  VERSION = "0.4.1"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gtfs_df
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 0.4.1
 platform: ruby
 authors:
 - David Mejorado