gtfs_df 0.11.0 → 0.12.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3258de8699df9b1e0eb934b3edab17a0381536dbdf7bde0275f375854392478d
-  data.tar.gz: c67a356c1fca9818b89a936dac816826f4a00b50b990d352d59d96d1231a4104
+  metadata.gz: 8c1060a2f120fc6f620adc523d8038a05ec6896826fbb950973c14b8a406fc37
+  data.tar.gz: 6f75be7bbec96e5f75994fa130f9d599061d796ef702cad6c833d8edb873ef6b
 SHA512:
-  metadata.gz: 881cf6027f5e68eda3916153889ac96f27f992ac411564a5232ae6b409b2488d41b596bcc23a6a4492c0a9fc2fce471986dd5ce07ede354a87a4b21033e0d4c8
-  data.tar.gz: 6f68989b3220ba20c6fc185560f0fedda797d4f0e06735a3c366f1722ecf6ce9fa35698716c0f339852202ceda7b9f4c8ef13fcb95b3e185a83202755d84f00e
+  metadata.gz: c8000394c34a67284d16df01e9e236b39d074bd300d1416be8557a7db2387357e860b7c741ff3eb483293895ca16bb3936efd857489124d309ced30a06d248aa
+  data.tar.gz: 15a5d19bb1d06ab1c9f21ae0e403ce6187dd370dcf868d20b9ee1de415a2a1ecabb1ae158b5dc10bc6236f329863e661581a2f8b39d0fb9eed50bc9dbad350db

data/CHANGELOG.md

@@ -1,8 +1,33 @@
+## [0.12.0] - 2026-06-12
+
+### 🚀 Features
+
+- Allow loading a subset of tables
+
+### 🐛 Bug Fixes
+
+- Truncate ragged lines
+- Consider BOM when looking for empty files
+## [0.11.1] - 2026-04-28
+
+### 🐛 Bug Fixes
+
+- Account for unreferenced calendar entries
+
+### 🧪 Testing
+
+- Validate filtering from feed with empty calendar
+- Validate unreferenced calendar dates
+
+### ⚙️ Miscellaneous Tasks
+
+- Bump version to 0.11.1
 ## [0.11.0] - 2026-03-30
 
 ### ⚙️ Miscellaneous Tasks
 
 - [**breaking**] Update ruby-polars dependency
+- Bump version to 0.11.0
 ## [0.10.2] - 2026-03-20
 
 ### 🐛 Bug Fixes

data/lib/gtfs_df/base_gtfs_table.rb

@@ -12,7 +12,13 @@ module GtfsDf
         elsif input.is_a?(String)
           # 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, encoding: "utf8-lossy")
+          df = Polars
+            .read_csv(
+              input,
+              infer_schema_length: 0,
+              encoding: "utf8-lossy",
+              truncate_ragged_lines: true
+            )
             .rename(->(col) { col.strip })
 
           # Strip out empty lines. Unfortunately read_csv does not support the drop_empty_rows

data/lib/gtfs_df/feed.rb

@@ -432,21 +432,22 @@ module GtfsDf
               # union semantics across those two parents (a trip is valid if it appears in
               # either).
               #
-              # Here we accumulate valid service_ids across calendar/calendar_dates, but only
-              # within the pool of trips that are already reachable from structural parents.
+              # Accumulate service_ids from each calendar source, then apply the filter.
+              # If the filter results in 0 trips, we continue accumulating to allow the next
+              # calendar edge to add valid service_ids. This handles feeds where
+              # calendar.txt has unreferenced service_ids but all trips use
+              # calendar_dates.txt service_ids.
               accumulated_service_ids = Polars.concat([accumulated_service_ids, valid_values]).unique
-
-              # Determine the base pool of trips:
-              # - If we've already restricted trips via structural parents (routes,
-              #   stop_times, shapes, etc), use that as the base.
-              # - Otherwise, like when filtering directly on trips, use the current
-              #   trips dataframe.
               trips_base_df ||= filtered[child_node.fetch(:file)]
               next unless trips_base_df && trips_base_df.height > 0
 
-              filtered[child_node.fetch(:file)] = trips_base_df.filter(
+              filtered_trips = trips_base_df.filter(
                 Polars.col("service_id").is_in(accumulated_service_ids.implode)
               )
+
+              if filtered_trips.height > 0
+                filtered[child_node.fetch(:file)] = filtered_trips
+              end
             else
               # Original single-edge logic for all other nodes
               before = child_df.height

data/lib/gtfs_df/reader.rb

@@ -6,13 +6,13 @@ module GtfsDf
     #
     # @param zip_path [String] Path to the GTFS zip file
     # @param parse_times [Boolean] Whether to parse time fields to seconds since midnight (default: false)
+    # @param relevant_files [Array<String>] A list of file names, useful to avoid loading tables you don't care about.
     # @return [Feed] The loaded GTFS feed
-    def self.load_from_zip(zip_path, parse_times: false)
+    def self.load_from_zip(zip_path, parse_times: false, relevant_files: nil)
       data = nil
 
-      relevant_files = GtfsDf::Feed::GTFS_FILES
-        .map { |name| "#{name}.txt" }
-        .to_set
+      relevant_files ||= GtfsDf::Feed::GTFS_FILES.map { |name| "#{name}.txt" }
+      relevant_files = relevant_files.to_set
 
       seen = {}
 
@@ -37,7 +37,7 @@ module GtfsDf
           end
         end
 
-        data = load_from_dir(tmpdir, parse_times: parse_times)
+        data = load_from_dir(tmpdir, parse_times:, relevant_files:)
       end
 
       data
@@ -47,12 +47,17 @@ module GtfsDf
     #
     # @param dir_path [String] Path to the GTFS directory
     # @param parse_times [Boolean] Whether to parse time fields to seconds since midnight (default: false)
+    # @param relevant_files [Array<String>] A list of file names, useful to avoid loading tables you don't care about.
     # @return [Feed] The loaded GTFS feed
-    def self.load_from_dir(dir_path, parse_times: false)
+    def self.load_from_dir(dir_path, parse_times: false, relevant_files: nil)
+      relevant_files ||= GtfsDf::Feed::GTFS_FILES.map { |name| "#{name}.txt" }
+      relevant_files = relevant_files.to_set
+
       data = {}
       GtfsDf::Feed::GTFS_FILES.each do |gtfs_file|
-        path = File.join(dir_path, "#{gtfs_file}.txt")
-        next unless File.exist?(path)
+        basename = "#{gtfs_file}.txt"
+        path = File.join(dir_path, basename)
+        next unless relevant_files.include?(basename) && File.exist?(path)
 
         data[gtfs_file] = data_frame(gtfs_file, path)
       end
@@ -66,7 +71,11 @@ module GtfsDf
     end
 
     private_class_method def self.has_header?(zip_entry)
-      zip_entry.get_input_stream.readline.strip != ""
+      zip_entry
+        .get_input_stream
+        .readline
+        .delete_prefix("\xEF\xBB\xBF".b) # BOM
+        .strip != ""
     rescue
       false
     end

data/lib/gtfs_df/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module GtfsDf
-  VERSION = "0.11.0"
+  VERSION = "0.12.0"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gtfs_df
 version: !ruby/object:Gem::Version
-  version: 0.11.0
+  version: 0.12.0
 platform: ruby
 authors:
 - David Mejorado