gtfs_df 0.10.2 → 0.11.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f128bb120303c7d0a4508f0d036812375893ea8c210a351535199b26f3aa3263
-  data.tar.gz: 554926a20907e8dedf02f786daa422a6bce1e3f876074a6b8106682b8a671eeb
+  metadata.gz: 70a7d05d5600baac0b36b21d0ebdd8fd0719cfe572844e44ae710c25fbf353f8
+  data.tar.gz: df2b5f47cd19bfe213644dcda42b22778acf98d8f104bf71c4e231eaa4860bba
 SHA512:
-  metadata.gz: 140df9b1b30d5210e6212ee8c7ee158217d26e2605ce9da9d0502eec90777ec4a4ee06a81555d858d69b53614af41d153836532aadd1a0053ec539b8caa9337e
-  data.tar.gz: 9f4fa2528b9897cac3aefb2019b80068adb2fc7022fd66601a91de284b7888cefafc4f5b2ad3f773ab01bb69849f43231fb6ef08155d0bc40fe1bcae9be5a8b1
+  metadata.gz: ee94819792b5c1389783a7844ce23af40de13264c5c069e608ae9c2383254c3fdcdc09915f5c33348d204675e5e373f8c26b039ff97144be37721aa7f4bc1b49
+  data.tar.gz: 28f28bb69a62738c817dafbc36c756a96f56f9f96d4d587aab2b9dffa8df5b6948b25d6b024e70856639cc2a97c75c49ff510bb7695c825f54405bf23701582d

data/CHANGELOG.md

@@ -1,3 +1,19 @@
+## [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
+## [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
@@ -7,6 +23,7 @@
 ### ⚙️ Miscellaneous Tasks
 
 - Raise clear message for multi-feed inputs
+- Bump version to 0.10.2
 ## [0.10.1] - 2026-03-18
 
 ### 🐛 Bug Fixes

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)))
@@ -430,21 +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/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 = "0.10.2"
+  VERSION = "0.11.1"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gtfs_df
 version: !ruby/object:Gem::Version
-  version: 0.10.2
+  version: 0.11.1
 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