philiprehberger-csv_kit 0.6.0 → 0.8.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7a49695d1e89645d30e0d325f7c7967db04b044f0cfe69ecf66cabfcc7e28470
-  data.tar.gz: 8a0cd35cd652e1893493ac4a0428251517c4a234cc0dca7b1addc9580b3688c6
+  metadata.gz: c48c099081a08bc9be83c75fb8879f9cb876ce2cc6e8978a44b8c18d3cd43776
+  data.tar.gz: fe72a612c3a6e5b0166c5eb24694f5593cbb0baca3eb541081dc84414939019b
 SHA512:
-  metadata.gz: 528761ab2269e586d1d01c20885b5113dd12fe89b94b266b4bcb7452224cedd19a626100218d77def4db8b4d087117717dd3a6738aa5fe54c4c05f20b7592b5c
-  data.tar.gz: 2838f1cb2c0bb9caa43015b2f8f07e3b6252f80d4ad376160619d2c3dc50c4f1fecb65d6d5c47f9e86ffa2537e6b9ace01623a266aa2bc9cfd0fd4c0e9977512
+  metadata.gz: bfff72e557a3deef83f104118ec4dddf00e5855b2d29ea1ac09a023f767a3b557f7415a9b8fd953ddca68a5cc08155b4b37b14a1800b3f2672db059b1a8a02da
+  data.tar.gz: a2bd950d995457677e70c2398efb4decb742597bc032e5f0f0454ac5a9025d4a41c2d05f4bcf18988f506c46ffdd0b20b354b7620748587c3de7e32da75387d7

data/CHANGELOG.md

@@ -7,6 +7,17 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.8.0] - 2026-04-17
+
+### Added
+- `CsvKit.to_csv(rows, headers:, dialect:)` — serialize an array of hashes to a CSV string; inverse of `to_hashes`
+- `to_hashes`, `pluck`, `headers`, `count`, `each_hash`, `find`, and `filter` now accept an IO object in addition to a file path
+
+## [0.7.0] - 2026-04-16
+
+### Added
+- `CsvKit.sample(path, n, dialect:)` — return n randomly sampled rows as symbolized hashes using reservoir sampling (Algorithm R); O(n) memory regardless of file size; returns all rows if file has fewer than n rows
+
 ## [0.6.0] - 2026-04-15
 
 ### Added
@@ -96,6 +107,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Type coercion and row validation
 - Quick load and filtering convenience methods
 
+[Unreleased]: https://github.com/philiprehberger/rb-csv-kit/compare/v0.8.0...HEAD
+[0.8.0]: https://github.com/philiprehberger/rb-csv-kit/compare/v0.7.0...v0.8.0
+[0.7.0]: https://github.com/philiprehberger/rb-csv-kit/compare/v0.6.0...v0.7.0
+[0.6.0]: https://github.com/philiprehberger/rb-csv-kit/compare/v0.5.0...v0.6.0
 [0.5.0]: https://github.com/philiprehberger/rb-csv-kit/releases/tag/v0.5.0
 [0.4.0]: https://github.com/philiprehberger/rb-csv-kit/releases/tag/v0.4.0
 [0.3.1]: https://github.com/philiprehberger/rb-csv-kit/releases/tag/v0.3.1

data/README.md

@@ -69,6 +69,15 @@ adults = Philiprehberger::CsvKit.each_hash("data.csv")
   .first(10)
 ```
 
+### Reservoir Sampling
+
+Return n randomly sampled rows with O(n) memory using Knuth's Algorithm R. If the file has fewer than n rows, all rows are returned:
+
+```ruby
+rows = Philiprehberger::CsvKit.sample("large.csv", 100)
+# => [{name: "Alice", age: "30"}, ...]
+```
+
 ### Find First Match
 
 Return the first row that matches a predicate, streaming and stopping on the first hit:
@@ -115,6 +124,21 @@ rows = Philiprehberger::CsvKit.process("data.csv", dialect: { delimiter: ";", qu
 end
 ```
 
+### Write CSV String
+
+Inverse of `to_hashes`. Serialize an array of hashes to a CSV string. Headers default to the keys of the first row:
+
+```ruby
+csv = Philiprehberger::CsvKit.to_csv([
+  { name: "Alice", age: 30 },
+  { name: "Bob",   age: 25 }
+])
+# => "name,age\nAlice,30\nBob,25\n"
+
+# Control column order / subset with explicit headers
+Philiprehberger::CsvKit.to_csv(rows, headers: [:name])
+```
+
 ### Writing CSV
 
 ```ruby
@@ -174,13 +198,15 @@ delimiter = Philiprehberger::CsvKit::Detector.detect("data.tsv")
 
 | Method / Class | Description |
 |----------------|-------------|
-| `CsvKit.to_hashes(path, dialect:)` | Load CSV into array of symbolized hashes |
-| `CsvKit.pluck(path, *keys, dialect:)` | Extract specific columns |
-| `CsvKit.filter(path, dialect:, &block)` | Filter rows, return CSV string |
-| `CsvKit.find(path, dialect:, &block)` | Return the first row matching the predicate, or nil |
-| `CsvKit.headers(path, dialect:)` | Return header row as array of symbols |
-| `CsvKit.count(path, dialect:)` | Count data rows without loading into memory |
-| `CsvKit.each_hash(path, dialect:, &block)` | Stream rows as symbolized hashes; returns Enumerator if no block |
+| `CsvKit.to_hashes(path_or_io, dialect:)` | Load CSV into array of symbolized hashes |
+| `CsvKit.to_csv(rows, headers:, dialect:)` | Serialize an array of hashes to a CSV string |
+| `CsvKit.sample(path_or_io, n, dialect:)` | Return n randomly sampled rows using reservoir sampling (Algorithm R) |
+| `CsvKit.pluck(path_or_io, *keys, dialect:)` | Extract specific columns |
+| `CsvKit.filter(path_or_io, dialect:, &block)` | Filter rows, return CSV string |
+| `CsvKit.find(path_or_io, dialect:, &block)` | Return the first row matching the predicate, or nil |
+| `CsvKit.headers(path_or_io, dialect:)` | Return header row as array of symbols |
+| `CsvKit.count(path_or_io, dialect:)` | Count data rows without loading into memory |
+| `CsvKit.each_hash(path_or_io, dialect:, &block)` | Stream rows as symbolized hashes; returns Enumerator if no block |
 | `CsvKit.process(path_or_io, dialect:, &block)` | Streaming DSL with transforms and validations |
 | `Processor#headers(*names)` | Override header names |
 | `Processor#transform(key, &block)` | Register column transform |

data/lib/philiprehberger/csv_kit/version.rb

@@ -2,6 +2,6 @@
 
 module Philiprehberger
   module CsvKit
-    VERSION = '0.6.0'
+    VERSION = '0.8.0'
   end
 end

data/lib/philiprehberger/csv_kit.rb

@@ -3,6 +3,7 @@
 require 'csv'
 require 'date'
 require 'time'
+require 'stringio'
 require_relative 'csv_kit/version'
 require_relative 'csv_kit/dialect'
 require_relative 'csv_kit/detector'
@@ -30,69 +31,85 @@ module Philiprehberger
 
     # Load an entire CSV into an array of symbolized hashes.
     #
-    # @param path [String] file path
+    # @param path_or_io [String, IO] file path or IO object
     # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
     # @return [Array<Hash{Symbol => String}>]
-    def self.to_hashes(path, dialect: nil)
-      csv_opts = { headers: true }
-      csv_opts = Dialect.new(dialect).merge_into(csv_opts) if dialect
-      CSV.foreach(path, **csv_opts).map do |row|
-        row.to_h.transform_keys(&:to_sym)
+    def self.to_hashes(path_or_io, dialect: nil)
+      rows = []
+      foreach_row(path_or_io, headers: true, dialect: dialect) do |row|
+        rows << row.to_h.transform_keys(&:to_sym)
+      end
+      rows
+    end
+
+    # Serialize an array of hashes to a CSV string.
+    #
+    # If headers is omitted, the keys of the first hash are used. Empty input
+    # returns an empty string. Dialect options are passed through to the writer.
+    #
+    # @param rows [Array<Hash>] data rows
+    # @param headers [Array<Symbol, String>, nil] explicit column order (optional)
+    # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
+    # @return [String] CSV string with header row
+    def self.to_csv(rows, headers: nil, dialect: nil)
+      return '' if rows.empty? && headers.nil?
+
+      resolved_headers = (headers || rows.first.keys).map(&:to_sym)
+      io = StringIO.new
+      Writer.stream(io, headers: resolved_headers, dialect: dialect) do |w|
+        rows.each { |row| w << (row.is_a?(Hash) ? row.transform_keys(&:to_sym) : row) }
       end
+      io.string
     end
 
     # Extract specific columns from a CSV.
     #
-    # @param path [String] file path
+    # @param path_or_io [String, IO] file path or IO object
     # @param keys [Array<Symbol>] column names to extract
     # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
     # @return [Array<Hash{Symbol => String}>]
-    def self.pluck(path, *keys, dialect: nil)
-      to_hashes(path, dialect: dialect).map { |h| h.slice(*keys) }
+    def self.pluck(path_or_io, *keys, dialect: nil)
+      to_hashes(path_or_io, dialect: dialect).map { |h| h.slice(*keys) }
     end
 
     # Return the header row as an array of symbols.
     #
-    # @param path [String] file path
+    # @param path_or_io [String, IO] file path or IO object
     # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
     # @return [Array<Symbol>]
-    def self.headers(path, dialect: nil)
+    def self.headers(path_or_io, dialect: nil)
       csv_opts = {}
       csv_opts = Dialect.new(dialect).merge_into(csv_opts) if dialect
-      CSV.open(path, **csv_opts) do |csv|
+      row = nil
+      with_csv(path_or_io, csv_opts) do |csv|
         row = csv.shift
-        return [] unless row
-
-        row.map(&:to_sym)
       end
+      return [] unless row
+
+      row.map(&:to_sym)
     end
 
     # Count data rows without loading them all into memory.
     #
-    # @param path [String] file path
+    # @param path_or_io [String, IO] file path or IO object
     # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
     # @return [Integer]
-    def self.count(path, dialect: nil)
-      csv_opts = { headers: true }
-      csv_opts = Dialect.new(dialect).merge_into(csv_opts) if dialect
+    def self.count(path_or_io, dialect: nil)
       n = 0
-      CSV.foreach(path, **csv_opts) { |_| n += 1 }
+      foreach_row(path_or_io, headers: true, dialect: dialect) { |_| n += 1 }
       n
     end
 
     # Stream rows one at a time as symbolized hashes with constant memory.
     # Returns an Enumerator if no block is given.
     #
-    # @param path [String] file path
+    # @param path_or_io [String, IO] file path or IO object
     # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
     # @yield [Hash{Symbol => String}] each row
     # @return [Enumerator, nil]
-    def self.each_hash(path, dialect: nil, &block)
-      csv_opts = { headers: true }
-      csv_opts = Dialect.new(dialect).merge_into(csv_opts) if dialect
-
+    def self.each_hash(path_or_io, dialect: nil, &block)
       enum = Enumerator.new do |yielder|
-        CSV.foreach(path, **csv_opts) do |row|
+        foreach_row(path_or_io, headers: true, dialect: dialect) do |row|
           yielder.yield(row.to_h.transform_keys(&:to_sym))
         end
       end
@@ -100,16 +117,40 @@ module Philiprehberger
       block ? enum.each(&block) : enum
     end
 
+    # Return n randomly sampled rows using reservoir sampling (Algorithm R).
+    # Memory usage is O(n) regardless of file size.
+    # If the file has fewer than n rows, all rows are returned.
+    #
+    # @param path_or_io [String, IO] file path or IO object
+    # @param n [Integer] number of rows to sample
+    # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
+    # @return [Array<Hash{Symbol => String}>]
+    def self.sample(path_or_io, n, dialect: nil)
+      reservoir = []
+      index = 0
+
+      foreach_row(path_or_io, headers: true, dialect: dialect) do |row|
+        hash = row.to_h.transform_keys(&:to_sym)
+        if index < n
+          reservoir << hash
+        else
+          j = rand(index + 1)
+          reservoir[j] = hash if j < n
+        end
+        index += 1
+      end
+
+      reservoir
+    end
+
     # Find the first row matching a predicate, streaming (stops as soon as a match is found).
     #
-    # @param path [String] file path
+    # @param path_or_io [String, IO] file path or IO object
     # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
     # @yield [Hash{Symbol => String}] each row as a symbolized hash
     # @return [Hash{Symbol => String}, nil] the first matching row or nil
-    def self.find(path, dialect: nil, &block)
-      csv_opts = { headers: true }
-      csv_opts = Dialect.new(dialect).merge_into(csv_opts) if dialect
-      CSV.foreach(path, **csv_opts) do |row|
+    def self.find(path_or_io, dialect: nil, &block)
+      foreach_row(path_or_io, headers: true, dialect: dialect) do |row|
         hash = row.to_h.transform_keys(&:to_sym)
         return hash if block.call(hash)
       end
@@ -118,12 +159,12 @@ module Philiprehberger
 
     # Filter rows and return matching rows as a CSV string.
     #
-    # @param path [String] file path
+    # @param path_or_io [String, IO] file path or IO object
     # @param dialect [Symbol, Hash, nil] CSV dialect preset or custom options
     # @yield [Hash{Symbol => String}] each row as a symbolized hash
     # @return [String] CSV string with headers
-    def self.filter(path, dialect: nil, &)
-      rows = to_hashes(path, dialect: dialect).select(&)
+    def self.filter(path_or_io, dialect: nil, &)
+      rows = to_hashes(path_or_io, dialect: dialect).select(&)
       return '' if rows.empty?
 
       headers = rows.first.keys
@@ -132,5 +173,29 @@ module Philiprehberger
         rows.each { |row| csv << headers.map { |k| row[k] } }
       end
     end
+
+    # @api private
+    # Iterate CSV rows from either a file path or an IO object.
+    def self.foreach_row(path_or_io, headers: false, dialect: nil, &block)
+      csv_opts = headers ? { headers: true } : {}
+      csv_opts = Dialect.new(dialect).merge_into(csv_opts) if dialect
+      if path_or_io.is_a?(String)
+        CSV.foreach(path_or_io, **csv_opts, &block)
+      else
+        CSV.new(path_or_io, **csv_opts).each(&block)
+      end
+    end
+
+    # @api private
+    # Open a CSV reader over either a file path or an IO object.
+    def self.with_csv(path_or_io, csv_opts, &block)
+      if path_or_io.is_a?(String)
+        CSV.open(path_or_io, **csv_opts, &block)
+      else
+        block.call(CSV.new(path_or_io, **csv_opts))
+      end
+    end
+
+    private_class_method :foreach_row, :with_csv
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: philiprehberger-csv_kit
 version: !ruby/object:Gem::Version
-  version: 0.6.0
+  version: 0.8.0
 platform: ruby
 authors:
 - Philip Rehberger
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-04-15 00:00:00.000000000 Z
+date: 2026-04-17 00:00:00.000000000 Z
 dependencies: []
 description: Streaming CSV processor with row-by-row transforms, validations, column
   plucking, streaming each_hash iteration, filtering, writing, error recovery, and