jklamer commented on code in PR #1631: URL: https://github.com/apache/avro/pull/1631#discussion_r846570543
########## lang/rust/avro_derive/src/lib.rs: ########## @@ -0,0 +1,366 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, +// software distributed under the License is distributed on an +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +// KIND, either express or implied. See the License for the +// specific language governing permissions and limitations +// under the License. + +use proc_macro2::{Span, TokenStream, TokenTree}; +use quote::quote; + +use syn::{parse_macro_input, Attribute, DeriveInput, Error, Lit, Path, Type, TypePath}; + +#[proc_macro_derive(AvroSchema, attributes(namespace))] +// Templated from Serde +pub fn proc_macro_derive_avro_schema(input: proc_macro::TokenStream) -> proc_macro::TokenStream { + let mut input = parse_macro_input!(input as DeriveInput); + derive_avro_schema(&mut input) + .unwrap_or_else(to_compile_errors) + .into() +} + +fn derive_avro_schema(input: &mut DeriveInput) -> Result<TokenStream, Vec<syn::Error>> { + let namespace = get_namespace_from_attributes(&input.attrs)?; + let full_schema_name = vec![namespace, Some(input.ident.to_string())] + .into_iter() + .flatten() + .collect::<Vec<String>>() + .join("."); + let schema_def = match &input.data { + syn::Data::Struct(s) => { + get_data_struct_schema_def(&full_schema_name, s, input.ident.span())? + } + syn::Data::Enum(e) => get_data_enum_schema_def(&full_schema_name, e, input.ident.span())?, + _ => { + return Err(vec![Error::new( + input.ident.span(), + "AvroSchema derive only works for structs and simple enums ", + )]) + } + }; + + let ty = &input.ident; + let (impl_generics, ty_generics, where_clause) = input.generics.split_for_impl(); + Ok(quote! { + impl #impl_generics apache_avro::schema::AvroSchemaWithResolved for #ty #ty_generics #where_clause { + fn get_schema_with_resolved(resolved_schemas: &mut HashMap<apache_avro::schema::Name, apache_avro::schema::Schema>) -> apache_avro::schema::Schema { + let name = apache_avro::schema::Name::new(#full_schema_name).expect(&format!("Unable to parse schema name {}", #full_schema_name)[..]); + if resolved_schemas.contains_key(&name) { + resolved_schemas.get(&name).unwrap().clone() + }else { + resolved_schemas.insert(name.clone(), Schema::Ref{name: name.clone()}); + #schema_def + } + } + } + }) +} + +fn get_namespace_from_attributes(attrs: &[Attribute]) -> Result<Option<String>, Vec<Error>> { + let namespace_attr_path_constant: Path = syn::parse2::<Path>(quote! {namespace}).unwrap(); + const NAMESPACE_PARSING_ERROR_CONSTANST: &str = + "Namespace attribute must be in form #[namespace = \"com.testing.namespace\"]"; + // parse out namespace if present. Requires strict syntax + for attr in attrs { + if namespace_attr_path_constant == attr.path { + let mut input_tokens = attr.tokens.clone().into_iter(); + if let ( + Some(TokenTree::Punct(punct)), + Some(TokenTree::Literal(namespace_literal)), + None, + ) = ( + input_tokens.next(), + input_tokens.next(), + input_tokens.next(), + ) { + if punct.as_char() == '=' { + if let Lit::Str(lit_str) = Lit::new(namespace_literal) { + return Ok(Some(lit_str.value())); + } + } + } + return Err(vec![Error::new_spanned( + &attr.tokens, + NAMESPACE_PARSING_ERROR_CONSTANST, + )]); + } + } + Ok(None) +} + +fn get_data_struct_schema_def( + full_schema_name: &str, + s: &syn::DataStruct, + error_span: Span, +) -> Result<TokenStream, Vec<Error>> { + let mut record_field_exprs = vec![]; + match s.fields { + syn::Fields::Named(ref a) => { + for (position, field) in a.named.iter().enumerate() { + let name = field.ident.as_ref().unwrap().to_string(); // we know everything has a name + let schema_expr = type_to_schema_expr(&field.ty)?; + let position = position; + record_field_exprs.push(quote! { + apache_avro::schema::RecordField { + name: #name.to_string(), + doc: Option::None, + default: Option::None, + schema: #schema_expr, + order: apache_avro::schema::RecordFieldOrder::Ignore, + position: #position, + } + }); + } + } + syn::Fields::Unnamed(_) => { + return Err(vec![Error::new( + error_span, + "AvroSchema derive does not work for tuple structs", + )]) + } + syn::Fields::Unit => { + return Err(vec![Error::new( + error_span, + "AvroSchema derive does not work for unit structs", + )]) + } + } + Ok(quote! { + let schema_fields = vec![#(#record_field_exprs),*]; + let name = apache_avro::schema::Name::new(#full_schema_name).expect(&format!("Unable to struct name for schema {}", #full_schema_name)[..]); + apache_avro::schema::record_schema_for_fields(name, None, None, schema_fields) + }) +} + +fn get_data_enum_schema_def( + full_schema_name: &str, + e: &syn::DataEnum, + error_span: Span, +) -> Result<TokenStream, Vec<Error>> { + if e.variants.iter().all(|v| syn::Fields::Unit == v.fields) { + let symbols: Vec<String> = e + .variants + .iter() + .map(|varient| varient.ident.to_string()) + .collect(); + Ok(quote! { + apache_avro::schema::Schema::Enum { + name: apache_avro::schema::Name::new(#full_schema_name).expect(&format!("Unable to parse enum name for schema {}", #full_schema_name)[..]), + aliases: None, + doc: None, + symbols: vec![#(#symbols.to_owned()),*] + } + }) + } else { + Err(vec![Error::new( + error_span, + "AvroSchema derive does not work for enums with non unit structs", + )]) + } +} + +/// Takes in the Tokens of a type and returns the tokens of an expression with return type `Schema` +fn type_to_schema_expr(ty: &Type) -> Result<TokenStream, Vec<Error>> { + if let Type::Path(p) = ty { + let type_string = p.path.segments.last().unwrap().ident.to_string(); + + let schema = match &type_string[..] { + "bool" => quote! {Schema::Boolean}, + "i8" | "i16" | "i32" | "u8" | "u16" => quote! {apache_avro::schema::Schema::Int}, + "i64" => quote! {apache_avro::schema::Schema::Long}, + "f32" => quote! {apache_avro::schema::Schema::Float}, + "f64" => quote! {apache_avro::schema::Schema::Double}, + "String" | "str" => quote! {apache_avro::schema::Schema::String}, + "char" => { + return Err(vec![Error::new_spanned( Review Comment: great question. I think I need to remove that to remain consistent with "Only derive schema for types we know we can both serialize and deserialize using the serde way" -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
