Make Puffin APIs public

Author: fqaiser94

crates/iceberg/src/lib.rs

@@ -88,4 +88,4 @@ mod utils;
 pub mod writer;
 
 mod delete_vector;
-mod puffin;
+pub mod puffin;

crates/iceberg/src/puffin/blob.rs

@@ -18,21 +18,21 @@
 use std::collections::HashMap;
 
 /// A serialized form of a "compact" Theta sketch produced by the Apache DataSketches library.
-pub(crate) const APACHE_DATASKETCHES_THETA_V1: &str = "apache-datasketches-theta-v1";
+pub const APACHE_DATASKETCHES_THETA_V1: &str = "apache-datasketches-theta-v1";
 
 /// The blob
 #[derive(Debug, PartialEq, Clone)]
-pub(crate) struct Blob {
+pub struct Blob {
     /// See blob types: https://iceberg.apache.org/puffin-spec/#blob-types
-    pub(crate) r#type: String,
+    pub r#type: String,
     /// List of field IDs the blob was computed for; the order of items is used to compute sketches stored in the blob.
-    pub(crate) fields: Vec<i32>,
+    pub fields: Vec<i32>,
     /// ID of the Iceberg table's snapshot the blob was computed from
-    pub(crate) snapshot_id: i64,
+    pub snapshot_id: i64,
     /// Sequence number of the Iceberg table's snapshot the blob was computed from
-    pub(crate) sequence_number: i64,
+    pub sequence_number: i64,
     /// The uncompressed blob data
-    pub(crate) data: Vec<u8>,
+    pub data: Vec<u8>,
     /// Arbitrary meta-information about the blob
-    pub(crate) properties: HashMap<String, String>,
+    pub properties: HashMap<String, String>,
 }

crates/iceberg/src/puffin/metadata.rs

@@ -26,33 +26,33 @@ use crate::{Error, ErrorKind, Result};
 
 /// Human-readable identification of the application writing the file, along with its version.
 /// Example: "Trino version 381"
-pub(crate) const CREATED_BY_PROPERTY: &str = "created-by";
+pub const CREATED_BY_PROPERTY: &str = "created-by";
 
 /// Metadata about a blob.
 /// For more information, see: https://iceberg.apache.org/puffin-spec/#blobmetadata
 #[derive(Debug, PartialEq, Eq, Serialize, Deserialize, Clone)]
 #[serde(rename_all = "kebab-case")]
-pub(crate) struct BlobMetadata {
+pub struct BlobMetadata {
     /// See blob types: https://iceberg.apache.org/puffin-spec/#blob-types
-    pub(crate) r#type: String,
+    pub r#type: String,
     /// List of field IDs the blob was computed for; the order of items is used to compute sketches stored in the blob.
-    pub(crate) fields: Vec<i32>,
+    pub fields: Vec<i32>,
     /// ID of the Iceberg table's snapshot the blob was computed from
-    pub(crate) snapshot_id: i64,
+    pub snapshot_id: i64,
     /// Sequence number of the Iceberg table's snapshot the blob was computed from
-    pub(crate) sequence_number: i64,
+    pub sequence_number: i64,
     /// The offset in the file where the blob contents start
-    pub(crate) offset: u64,
+    pub offset: u64,
     /// The length of the blob stored in the file (after compression, if compressed)
-    pub(crate) length: u64,
+    pub length: u64,
     /// The compression codec used to compress the data
     #[serde(skip_serializing_if = "CompressionCodec::is_none")]
     #[serde(default)]
-    pub(crate) compression_codec: CompressionCodec,
+    pub compression_codec: CompressionCodec,
     /// Arbitrary meta-information about the blob
     #[serde(skip_serializing_if = "HashMap::is_empty")]
     #[serde(default)]
-    pub(crate) properties: HashMap<String, String>,
+    pub properties: HashMap<String, String>,
 }
 
 #[derive(Clone, Copy, PartialEq, Eq, Hash, Debug)]
@@ -91,13 +91,13 @@ impl Flag {
 /// Metadata about a puffin file.
 /// For more information, see: https://iceberg.apache.org/puffin-spec/#filemetadata
 #[derive(Debug, PartialEq, Eq, Serialize, Deserialize, Clone)]
-pub(crate) struct FileMetadata {
+pub struct FileMetadata {
     /// Metadata about blobs in file
-    pub(crate) blobs: Vec<BlobMetadata>,
+    pub blobs: Vec<BlobMetadata>,
     /// Arbitrary meta-information, like writer identification/version.
     #[serde(skip_serializing_if = "HashMap::is_empty")]
     #[serde(default)]
-    pub(crate) properties: HashMap<String, String>,
+    pub properties: HashMap<String, String>,
 }
 
 impl FileMetadata {

crates/iceberg/src/puffin/mod.rs

@@ -18,15 +18,23 @@
 //! Iceberg Puffin implementation.
 
 #![deny(missing_docs)]
-// Temporarily allowing this while crate is under active development
-#![allow(dead_code)]
 
 mod blob;
+pub use blob::{Blob, APACHE_DATASKETCHES_THETA_V1};
+
 mod compression;
+pub use compression::CompressionCodec;
+
 mod metadata;
+pub use metadata::{BlobMetadata, FileMetadata, CREATED_BY_PROPERTY};
+
 #[cfg(feature = "tokio")]
 mod reader;
+#[cfg(feature = "tokio")]
+pub use reader::PuffinReader;
+
 mod writer;
+pub use writer::PuffinWriter;
 
 #[cfg(test)]
 mod test_utils;

crates/iceberg/src/puffin/reader.rs

@@ -23,29 +23,29 @@ use crate::puffin::metadata::{BlobMetadata, FileMetadata};
 use crate::Result;
 
 /// Puffin reader
-pub(crate) struct PuffinReader {
+pub struct PuffinReader {
     input_file: InputFile,
     file_metadata: OnceCell<FileMetadata>,
 }
 
 impl PuffinReader {
     /// Returns a new Puffin reader
-    pub(crate) fn new(input_file: InputFile) -> Self {
+    pub fn new(input_file: InputFile) -> Self {
         Self {
             input_file,
             file_metadata: OnceCell::new(),
         }
     }
 
     /// Returns file metadata
-    pub(crate) async fn file_metadata(&self) -> Result<&FileMetadata> {
+    pub async fn file_metadata(&self) -> Result<&FileMetadata> {
         self.file_metadata
             .get_or_try_init(|| FileMetadata::read(&self.input_file))
             .await
     }
 
     /// Returns blob
-    pub(crate) async fn blob(&self, blob_metadata: &BlobMetadata) -> Result<Blob> {
+    pub async fn blob(&self, blob_metadata: &BlobMetadata) -> Result<Blob> {
         let file_read = self.input_file.reader().await?;
         let start = blob_metadata.offset;
         let end = start + blob_metadata.length;

crates/iceberg/src/puffin/writer.rs

@@ -26,7 +26,7 @@ use crate::puffin::metadata::{BlobMetadata, FileMetadata, Flag};
 use crate::Result;
 
 /// Puffin writer
-pub(crate) struct PuffinWriter {
+pub struct PuffinWriter {
     writer: Box<dyn FileWrite>,
     is_header_written: bool,
     num_bytes_written: u64,
@@ -38,7 +38,7 @@ pub(crate) struct PuffinWriter {
 
 impl PuffinWriter {
     /// Returns a new Puffin writer
-    pub(crate) async fn new(
+    pub async fn new(
         output_file: &OutputFile,
         properties: HashMap<String, String>,
         compress_footer: bool,
@@ -63,11 +63,7 @@ impl PuffinWriter {
     }
 
     /// Adds blob to Puffin file
-    pub(crate) async fn add(
-        &mut self,
-        blob: Blob,
-        compression_codec: CompressionCodec,
-    ) -> Result<()> {
+    pub async fn add(&mut self, blob: Blob, compression_codec: CompressionCodec) -> Result<()> {
         self.write_header_once().await?;
 
         let offset = self.num_bytes_written;
@@ -89,7 +85,7 @@ impl PuffinWriter {
     }
 
     /// Finalizes the Puffin file
-    pub(crate) async fn close(mut self) -> Result<()> {
+    pub async fn close(mut self) -> Result<()> {
         self.write_header_once().await?;
         self.write_footer().await?;
         self.writer.close().await?;