parquet 0.3.2 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 259f41f1ea1b111a0b0fdec15d17b54ac5a4efee750157159a33a4272b0b1310
-  data.tar.gz: 71579d25b3ec411208103e4afe116285bd75a30003af3f21b7d7972ef8942ff6
+  metadata.gz: bb28e49c647fcf9dddba5d18806b366e73ecaa6fb3bf8a3534eb7df4749710f0
+  data.tar.gz: 44953fd9a1f3fc89f24754a6cb555f4d32a90d4af4bb4aa3c14fc373a1c389df
 SHA512:
-  metadata.gz: 792a6653554393a94de0572ad2164e5da89c7b49b9476599b9d45efd38448f6e7065fa5b7b0082e036ea19da19441b462083ca9e61226f82dd38a812905a189e
-  data.tar.gz: 1520073d0668751a5c449dde08deea5dd06e50845e8e8fead398d7015d5cc979885897a258f1ab86eb405203fcaf6ebe0164ccdb31e4033e4911ad4b68268d11
+  metadata.gz: eaa9767d45deab2c3ba3fb0fc4c7aeb222b4446494262ff01083bdc488cca3164feee0f2f07bff644a3196479da0db4049dee1395b8b89f87f5aeba24a97b60b
+  data.tar.gz: 805a9cad6698f6ade3a61b40eb514a948ed6b07682db0bd3e724aad533468a114053b9589ec6261c31e8fe6b244c4808457f9e801a6991dffa86a36326736ab0

data/ext/parquet/src/ruby_reader.rs

@@ -34,9 +34,10 @@ impl RubyReader {
     }
 
     // For now, don't use this. Having to use seek in length is scary.
-    fn is_seekable_io_like(_value: &Value) -> bool {
-        // Self::is_io_like(value) && value.respond_to("seek", false).unwrap_or(false)
-        false
+    fn is_seekable_io_like(value: &Value) -> bool {
+        Self::is_io_like(value)
+            && value.respond_to("seek", false).unwrap_or(false)
+            && value.respond_to("pos", false).unwrap_or(false)
     }
 }
 
@@ -92,10 +93,14 @@ impl Seek for RubyReader {
                 let unwrapped_inner = ruby.get_inner(*inner);
 
                 let new_offset = match pos {
-                    io::SeekFrom::Start(offset) => offset as usize,
-                    io::SeekFrom::Current(offset) => (*original_offset as i64 + offset) as usize,
-                    io::SeekFrom::End(offset) => {
-                        unwrapped_inner.len().saturating_sub(offset as usize)
+                    SeekFrom::Start(off) => off as usize,
+                    SeekFrom::Current(off) => {
+                        let signed = *original_offset as i64 + off;
+                        signed.max(0) as usize
+                    }
+                    SeekFrom::End(off) => {
+                        let signed = unwrapped_inner.len() as i64 + off;
+                        signed.max(0) as usize
                     }
                 };
 
@@ -111,8 +116,12 @@ impl Seek for RubyReader {
                     SeekFrom::End(i) => (2, i),
                 };
 
+                unwrapped_inner
+                    .funcall::<_, _, u64>("seek", (ruby_offset, whence))
+                    .map_err(|e| io::Error::new(io::ErrorKind::Other, e.to_string()))?;
+
                 let new_position = unwrapped_inner
-                    .funcall("seek", (ruby_offset, whence))
+                    .funcall::<_, _, u64>("pos", ())
                     .map_err(|e| io::Error::new(io::ErrorKind::Other, e.to_string()))?;
 
                 Ok(new_position)

data/ext/parquet/src/types/record_types.rs

@@ -136,7 +136,8 @@ impl TryIntoValue for ParquetField {
                         .map_err(|e| ReaderError::Utf8Error(e))
                         .and_then(|s| Ok(s.into_value_with(handle)))?)
                 } else {
-                    Ok(handle.str_from_slice(s.as_bytes()).as_value())
+                    let s = String::from_utf8_lossy(s.as_bytes());
+                    Ok(s.into_value_with(handle))
                 }
             }
             Field::Byte(b) => Ok(b.into_value_with(handle)),

data/ext/parquet/src/utils.rs

@@ -83,7 +83,7 @@ pub fn parse_parquet_rows_args(ruby: &Ruby, args: &[Value]) -> Result<ParquetRow
         None => ParserResultType::Hash,
     };
 
-    let strict = kwargs.optional.2.flatten().unwrap_or(false);
+    let strict = kwargs.optional.2.flatten().unwrap_or(true);
 
     Ok(ParquetRowsArgs {
         to_read,
@@ -159,6 +159,6 @@ pub fn parse_parquet_columns_args(
         result_type,
         columns: kwargs.optional.1.flatten(),
         batch_size: kwargs.optional.2.flatten(),
-        strict: kwargs.optional.3.flatten().unwrap_or(false),
+        strict: kwargs.optional.3.flatten().unwrap_or(true),
     })
 }

data/lib/parquet/version.rb

@@ -1,3 +1,3 @@
 module Parquet
-  VERSION = "0.3.2"
+  VERSION = "0.4.0"
 end

data/lib/parquet.rbi

@@ -11,7 +11,8 @@ module Parquet
     params(
       input: T.any(String, File, StringIO, IO),
       result_type: T.nilable(T.any(String, Symbol)),
-      columns: T.nilable(T::Array[String])
+      columns: T.nilable(T::Array[String]),
+      strict: T.nilable(T::Boolean)
     ).returns(T::Enumerator[T.any(T::Hash[String, T.untyped], T::Array[T.untyped])])
   end
   sig do
@@ -19,10 +20,11 @@ module Parquet
       input: T.any(String, File, StringIO, IO),
       result_type: T.nilable(T.any(String, Symbol)),
       columns: T.nilable(T::Array[String]),
+      strict: T.nilable(T::Boolean),
       blk: T.nilable(T.proc.params(row: T.any(T::Hash[String, T.untyped], T::Array[T.untyped])).void)
     ).returns(NilClass)
   end
-  def self.each_row(input, result_type: nil, columns: nil, &blk)
+  def self.each_row(input, result_type: nil, columns: nil, strict: nil, &blk)
   end
 
   # Options:
@@ -36,7 +38,8 @@ module Parquet
       input: T.any(String, File, StringIO, IO),
       result_type: T.nilable(T.any(String, Symbol)),
       columns: T.nilable(T::Array[String]),
-      batch_size: T.nilable(Integer)
+      batch_size: T.nilable(Integer),
+      strict: T.nilable(T::Boolean)
     ).returns(T::Enumerator[T.any(T::Hash[String, T.untyped], T::Array[T.untyped])])
   end
   sig do
@@ -45,11 +48,12 @@ module Parquet
       result_type: T.nilable(T.any(String, Symbol)),
       columns: T.nilable(T::Array[String]),
       batch_size: T.nilable(Integer),
+      strict: T.nilable(T::Boolean),
       blk:
         T.nilable(T.proc.params(batch: T.any(T::Hash[String, T::Array[T.untyped]], T::Array[T::Array[T.untyped]])).void)
     ).returns(NilClass)
   end
-  def self.each_column(input, result_type: nil, columns: nil, batch_size: nil, &blk)
+  def self.each_column(input, result_type: nil, columns: nil, batch_size: nil, strict: nil, &blk)
   end
 
   # Options:
@@ -80,7 +84,15 @@ module Parquet
       sample_size: T.nilable(Integer)
     ).void
   end
-  def self.write_rows(read_from, schema:, write_to:, batch_size: nil, flush_threshold: nil, compression: nil, sample_size: nil)
+  def self.write_rows(
+    read_from,
+    schema:,
+    write_to:,
+    batch_size: nil,
+    flush_threshold: nil,
+    compression: nil,
+    sample_size: nil
+  )
   end
 
   # Options:

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: parquet
 version: !ruby/object:Gem::Version
-  version: 0.3.2
+  version: 0.4.0
 platform: ruby
 authors:
 - Nathan Jaremko
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-02-19 00:00:00.000000000 Z
+date: 2025-02-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys