raka 0.8.2 → 0.9.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 106a319450b4f3049bf8c6aad08fe881408956b0501484a4c7307862a8fef68d
-  data.tar.gz: 863d3a79bdc3bb3b091401b72a167b180cf3cb8aaaf483a86a9bc2981f0985f2
+  metadata.gz: 8b04ff22d6ec2fd1a4c986b4a7969596ea9fbe155669558eb69dd469b947aebc
+  data.tar.gz: 680f48420d9d1f3bb0e758b21df43b97f7de9af489385b19fa465cc76bb3f816
 SHA512:
-  metadata.gz: f9a7de5feb97b7c55f6c2e31e0f51037a9e9842072738f695a41b086bb0564ac7fcf503d5f69061d5bdf2c82ddc0824bb3ffdf45c12386b4fe0efbac0d3c0f10
-  data.tar.gz: 8c1b81ff687f86dfb7dfde300365c12c4c06ec300b01748d89ea477cef85006bda86cbaa02945b7ad1f70ecb00d3df4f4c423dce4f28dbe96a9d8f425e41c653
+  metadata.gz: 41de921265fc6b0cd178405054dfed5b8c65acb99819f9fe1dfd1808d1a45e08dc19ef871e01566047f4014d1d0cae1235a9c2f0b71e8d67646b7a69d8e58826
+  data.tar.gz: f5c274b5a85941fd347822ce67e9aff14969c5e9816a7472c77005bc38c655e9ccb7e4d865319646e04b140fa14c3b2d72117d668066ef727084c2818a5c94f4

data/VERSION

@@ -1 +1 @@
-0.8.2
+0.9.0

data/lib/raka/lang/duckdb/impl.rb

@@ -17,12 +17,13 @@ end
 # 1. Persistent mode: operations on .db file with CREATE TABLE
 # 2. Ad-hoc mode: parquet in/out using COPY operations
 class Duckdb
-  def initialize(database: nil, params: {}, before: nil, after: nil)
+  def initialize(database: nil, params: {}, before: nil, after: nil, format: nil)
     @params = params
     @database = database
     @mode = @database ? :persistent : :adhoc
     @before = before
     @after = after
+    @format = format&.upcase
   end
 
   def duckdb_cmd
@@ -36,7 +37,7 @@ class Duckdb
 
   def process_params(code)
     return code if code.nil?
-    
+
     processed_code = code
     (@params || {}).each do |key, value|
       processed_code = processed_code.gsub("$#{key}", "'#{value}'")
@@ -44,7 +45,12 @@ class Duckdb
     processed_code
   end
 
-  def build(code, _task)
+  def detect_format_from_extension(filename)
+    ext = File.extname(filename)[1..]&.upcase  # Remove dot and convert to uppercase
+    ext || 'PARQUET'  # Default fallback
+  end
+
+  def build(code, task)
     # Process parameter placeholders for all parts
     main_sql = process_params(code)
     before_sql = process_params(@before)
@@ -52,22 +58,24 @@ class Duckdb
 
     # Build SQL parts as separate statements
     sql_parts = []
-    
+
     # Add before hook if present
     sql_parts << before_sql if before_sql
-    
+
     # Add main query based on mode
     case @mode
     when :persistent
-      sql_parts << "DROP TABLE IF EXISTS :_name_;"
+      sql_parts << 'DROP TABLE IF EXISTS :_name_;'
       sql_parts << "CREATE TABLE :_name_ AS (#{main_sql});"
     when :adhoc
-      sql_parts << "COPY (#{main_sql}) TO ':output:' (FORMAT PARQUET);"
+      # Determine format: use explicit format if provided, otherwise detect from output filename
+      format = @format || detect_format_from_extension(task.name)
+      sql_parts << "COPY (#{main_sql}) TO ':output:' (FORMAT #{format});"
     end
-    
+
     # Add after hook if present
     sql_parts << after_sql if after_sql
-    
+
     sql_parts.join("\n")
   end
 
@@ -77,12 +85,12 @@ class Duckdb
       # Split the SQL into separate statements and execute them individually
       bash env, %(
       # Execute the combined SQL script with proper variable replacement
-      cat #{fname} | sed 's|:_name_|#{task.output_stem}|g' | #{duckdb_cmd} | sed -z '$ s/\n$//' | tee #{fname}.log
+      cat #{fname} | sed 's|:_name_|#{task.output_stem}|g' | #{duckdb_cmd} | sed -z '$ s/\\n$//' | tee #{fname}.log
       echo "#{@database}" > #{task.name}
       )
     when :adhoc
       bash env, %(
-      cat #{fname} | sed 's|:output:|#{task.name}|g' | #{duckdb_cmd} | sed -z '$ s/\n$//' | tee #{fname}.log
+      cat #{fname} | sed 's|:output:|#{task.name}|g' | #{duckdb_cmd} | sed -z '$ s/\\n$//' | tee #{fname}.log
       )
     end
   end

data/lib/raka/lang/psql/impl.rb

@@ -52,7 +52,7 @@ class Psql
 
     bash env, %(
     #{sh_cmd(schema)} #{param_str} -v _name_=#{task.output_stem} \
-      -v _schema_=#{schema.empty? ? '' : schema + '.'} -f #{fname} | sed -z '$ s/\n$//' | tee #{fname}.log
+      -v _schema_=#{schema.empty? ? '' : schema + '.'} -f #{fname} | sed -z '$ s/\\n$//' | tee #{fname}.log
     mv #{fname}.log #{task.name}
     )
   end

data/lib/raka/protocol.rb

@@ -56,11 +56,11 @@ class LanguageProtocol
     code = yield @text if @text
     code = @block.call(task) if @block # do not resolve
 
-    env.logger.debug code
+    env.logger.debug code.chomp
     script_text = @impl.build(wrap_template(remove_common_indent(code)), task)
     temp_script = create_tmp(script_text)
     @impl.run_script env, temp_script, task
-    env.logger.debug script_text
+    env.logger.debug script_text.chomp
   end
 end
 
@@ -91,7 +91,7 @@ def run_cmd(env, cmd)
   if env.logger.level <= 0
     pid = spawn(cmd, out: out_w)
     Thread.new do
-      env.logger.debug(out_r.gets) until out_r.eof
+      env.logger.debug(out_r.gets.chomp) until out_r.eof
     end
   elsif env.logger.level == 1
     pid = spawn(cmd, out: out_w)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: raka
 version: !ruby/object:Gem::Version
-  version: 0.8.2
+  version: 0.9.0
 platform: ruby
 authors:
 - yarray