gtfs_df 0.10.1 → 0.11.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5200ee072c53d95cdb96af3fc8e0f15f9a15e2cc22684b5c9b2b2eeb4c103c8a
-  data.tar.gz: 843492a5f86c595d48431edcf2decd800debc1ead5dd2df0e77d89bf920c01ce
+  metadata.gz: 3258de8699df9b1e0eb934b3edab17a0381536dbdf7bde0275f375854392478d
+  data.tar.gz: c67a356c1fca9818b89a936dac816826f4a00b50b990d352d59d96d1231a4104
 SHA512:
-  metadata.gz: be450515810d21760afbff3800887e3e5605052bd52eaf7df81951d6193c141dc9487d6b566b9caae5bd44a990051f54907ded2fa619406018afb32df8f3209a
-  data.tar.gz: b6efebb9c5190777a3698a2b747b8528a7a72598c96254fbc7c9ead11aceaa72de75d2a95fb5a4c60745a0aacb0234108e4e5c663650bd5e0d71b858881325ca
+  metadata.gz: 881cf6027f5e68eda3916153889ac96f27f992ac411564a5232ae6b409b2488d41b596bcc23a6a4492c0a9fc2fce471986dd5ce07ede354a87a4b21033e0d4c8
+  data.tar.gz: 6f68989b3220ba20c6fc185560f0fedda797d4f0e06735a3c366f1722ecf6ce9fa35698716c0f339852202ceda7b9f4c8ef13fcb95b3e185a83202755d84f00e

data/CHANGELOG.md

@@ -1,3 +1,18 @@
+## [0.11.0] - 2026-03-30
+
+### ⚙️ Miscellaneous Tasks
+
+- [**breaking**] Update ruby-polars dependency
+## [0.10.2] - 2026-03-20
+
+### 🐛 Bug Fixes
+
+- Handle nested feeds
+
+### ⚙️ Miscellaneous Tasks
+
+- Raise clear message for multi-feed inputs
+- Bump version to 0.10.2
 ## [0.10.1] - 2026-03-18
 
 ### 🐛 Bug Fixes
@@ -8,6 +23,7 @@
 
 - Normalize platforms
 - Ignore the examples folder on publishing
+- Bump version to 0.10.1
 ## [0.10.0] - 2026-03-06
 
 ### 🚀 Features

data/lib/gtfs_df/base_gtfs_table.rb

@@ -22,7 +22,7 @@ module GtfsDf
           dtypes = self.class::SCHEMA.slice(*df.columns)
           df
             .with_columns(dtypes.keys.map do |col|
-              stripped = Polars.col(col).str.strip
+              stripped = Polars.col(col).str.strip_chars
               Polars.when(stripped.str.len_chars.gt(0))
                 .then(stripped)
                 .otherwise(Polars.lit(nil))

data/lib/gtfs_df/feed.rb

@@ -128,12 +128,12 @@ module GtfsDf
           trip_ids = if trip_ids.empty?
             new_filtered["trips"]["trip_id"]
           else
-            trip_ids.filter(trip_ids.is_in(new_filtered["trips"]["trip_id"]))
+            trip_ids.filter(trip_ids.is_in(new_filtered["trips"]["trip_id"].implode))
           end
         end
 
         if trip_ids
-          filtered = filter!("trips", {"trip_id" => trip_ids.implode}, filtered.dup)
+          filtered = filter!("trips", {"trip_id" => trip_ids}, filtered.dup)
         end
       end
 
@@ -330,7 +330,7 @@ module GtfsDf
 
       # Get the week with max trips
       # Sort by total_trips descending, then date ascending to pick the earliest date in case of a tie
-      sorted_weeks = weekly_agg.sort(["total_trips", "week_start"], reverse: [true, false])
+      sorted_weeks = weekly_agg.sort(["total_trips", "week_start"], descending: [true, false])
       best_week = sorted_weeks.head(1)
 
       return nil if best_week.height == 0
@@ -346,7 +346,9 @@ module GtfsDf
         df = filtered[file]
 
         filters.each do |col, val|
-          df = if val.is_a?(Polars::Series) || val.is_a?(Array)
+          df = if val.is_a?(Polars::Series)
+            df.filter(Polars.col(col).is_in(val.implode))
+          elsif val.is_a?(Array)
             df.filter(Polars.col(col).is_in(val))
           elsif val.respond_to?(:call)
             df.filter(val.call(Polars.col(col)))

data/lib/gtfs_df/reader.rb

@@ -14,15 +14,26 @@ module GtfsDf
         .map { |name| "#{name}.txt" }
         .to_set
 
+      seen = {}
+
       Dir.mktmpdir do |tmpdir|
         Zip::File.open(zip_path) do |zip_file|
           zip_file.each do |entry|
+            # Extract files in nested directories into the root of the tmpdir
+            file_name = File.basename(entry.name)
+
+            if seen[file_name]
+              raise GtfsDf::Error, "Found multiple instances of the same file: #{seen[file_name]} and #{entry.name}"
+            end
+
             # We're skipping:
-            # - directories
             # - unrelated files
             # - empty feed files
-            next unless entry.file? && relevant_files.include?(entry.name) && has_header?(entry)
-            entry.extract(destination_directory: tmpdir)
+            next unless relevant_files.include?(file_name) && has_header?(entry)
+
+            seen[file_name] = entry.name
+
+            entry.extract(file_name, destination_directory: tmpdir)
           end
         end
 

data/lib/gtfs_df/utils.rb

@@ -17,9 +17,9 @@ module GtfsDf
     def as_seconds_since_midnight(col_name)
       parts = Polars.col(col_name).str.split(":")
 
-      hours = parts.list.get(0).cast(:i64)
-      minutes = parts.list.get(1).cast(:i64)
-      seconds = parts.list.get(2).cast(:i64)
+      hours = parts.list.get(0).cast(Polars::Int64)
+      minutes = parts.list.get(1).cast(Polars::Int64)
+      seconds = parts.list.get(2).cast(Polars::Int64)
 
       (hours * SECONDS_IN_HOUR) +
         (minutes * SECONDS_IN_MINUTE) +
@@ -42,9 +42,9 @@ module GtfsDf
 
       Polars.format(
         "{}:{}:{}",
-        hours.cast(:str).str.zfill(2),
-        minutes.cast(:str).str.zfill(2),
-        seconds.cast(:str).str.zfill(2)
+        hours.cast(Polars::String).str.zfill(2),
+        minutes.cast(Polars::String).str.zfill(2),
+        seconds.cast(Polars::String).str.zfill(2)
       )
     end
 

data/lib/gtfs_df/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module GtfsDf
-  VERSION = "0.10.1"
+  VERSION = "0.11.0"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gtfs_df
 version: !ruby/object:Gem::Version
-  version: 0.10.1
+  version: 0.11.0
 platform: ruby
 authors:
 - David Mejorado
@@ -30,20 +30,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.22'
-    - - "<"
-      - !ruby/object:Gem::Version
-        version: '0.24'
+        version: '0.25'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.22'
-    - - "<"
-      - !ruby/object:Gem::Version
-        version: '0.24'
+        version: '0.25'
 - !ruby/object:Gem::Dependency
   name: rubyzip
   requirement: !ruby/object:Gem::Requirement