Merge branch 'development' into 'main'

Fixes mapping plan generation by handling edge cases around iri generations and term typing

See merge request rml/proc/algemaploom-rs!14

.gitignore

@@ -11,3 +11,4 @@ test_resources/
 
 node_modules/
 src/python/ltranslator.so
+.idea/

.gitlab-ci.yml

@@ -4,13 +4,31 @@ test:
   script:
     - cargo test --verbose --jobs 1
 
-build:
+build-no-bindings:
+  image: rustdocker/rust:stable
+  stage: build
+  script:
+    - cargo build
+
+build-java:
   image: rustdocker/rust:stable
   stage: build
   script:
     - apt update
     - apt install -y openjdk-17-jdk curl
-    - curl -fsSL https://deb.nodesource.com/setup_20.x | sudo bash - 
-    - sudo apt-get install -y nodejs
-    - ./build.sh
+    - ./build_java.sh
+
+build-python:
+  image: rustdocker/rust:stable
+  stage: build
+  script:
+    - ./build_python.sh
 
+build-nodejs:
+  image: rustdocker/rust:stable
+  stage: build
+  script:
+    - curl -fsSL https://deb.nodesource.com/setup_20.x | sudo bash -
+    - apt update
+    - sudo apt-get install -y nodejs
+    - ./build_nodejs.sh

.rustfmt.toml

@@ -0,0 +1,10 @@
+
+tab_spaces=4
+reorder_imports = true
+imports_granularity = "Module"
+group_imports = "StdExternalCrate"
+struct_field_align_threshold = 20
+use_field_init_shorthand = true
+enum_discrim_align_threshold = 20
+force_multiline_blocks = true
+max_width = 80

Cargo.toml

@@ -47,9 +47,9 @@ walkdir = "2.4.0"
 colored = "2.0.4"
 log4rs.workspace = true
 log.workspace = true
-jni = "0.21.1"
-neon = "1.0.0"
-pyo3 = { version = "0.21.2", features = ["extension-module"] }
+jni = { version = "0.21.1" , optional = true}
+neon = { version = "1.0.0", optional = true }
+pyo3 = { version = "0.21.2", features = ["extension-module"], optional = true }
 
 [lib]
 name = "ltranslator"

README.md

@@ -96,6 +96,14 @@ The following features are not supported in translation yet:
 4) Functions 
 5) Conditionals
 
+## Bindings
+AlgeMapLoom provides bindings for Java, Python and Node.js.
+These can be enabled with the features `jni`, `pyo3` and `neon` respectively.
+
+If you build from source, you can run the `build_java.sh`, `build_python.sh` and `build_nodejs.sh`
+scripts respectively.
+
+For usage, check out the [src/java](src/java), [src/python](src/python), and [src/nodejs](src/nodejs) folders.
 
 
 ## Acknowledgement

build_java.sh

@@ -0,0 +1,12 @@
+#!/usr/bin/env bash
+
+set -e
+
+# Rust build
+echo "==> Rust building and bindings"
+cargo build --lib --release --features=jni
+
+# Java bindings
+echo "==> Java bindings"
+# Compile Translator CLI
+javac src/java/Translator.java

build_nodejs.sh

@@ -0,0 +1,16 @@
+#!/usr/bin/env bash
+
+set -e
+
+# Rust build
+echo "==> Rust building and bindings"
+cargo build --lib --release --features=neon
+
+# NodeJS bindings
+echo "==> NodeJS bindings"
+# Install NodeJS NEON dependencies
+npm i
+# Execute NEON on Rust library to generate index.node file for NodeJS dynamic library
+./node_modules/.bin/neon dist -n translator -v -f target/release/libltranslator.so
+# Move index.node to the right folder after generation
+mv index.node src/nodejs

build_python.sh

@@ -0,0 +1,12 @@
+#!/usr/bin/env bash
+
+set -e
+
+# Rust build
+echo "==> Rust building and bindings"
+cargo build --lib --release --features=pyo3
+
+# Python bindings
+echo "==> Python bindings"
+# Native import, but requires renaming
+cp target/release/libltranslator.so src/python/ltranslator.so

operator/src/lib.rs

@@ -307,6 +307,7 @@ pub enum Function {
         inner_function: RcExtendFunction,
     },
     Iri {
+        base_iri: Option<String>, 
         inner_function: RcExtendFunction,
     },
     Literal {

rml-interpreter/src/extractors/io.rs

@@ -1,5 +1,5 @@
 use std::fs::File;
-use std::io::{BufRead, BufReader};
+use std::io::{BufRead, BufReader, Read, Seek};
 use std::path::PathBuf;
 
 use sophia_api::triple::stream::TripleSource;
@@ -11,6 +11,10 @@ use super::triplesmap_extractor::extract_triples_maps;
 use super::ExtractorResult;
 use crate::rml_model::Document;
 
+fn extract_base_iri(input: &str) -> Option<String> {
+    input.strip_prefix("@base").map(|e| e[0..e.len()-1].replace(['<', '>'], "").trim().to_string())
+}
+
 pub fn load_graph_bread(buf_read: impl BufRead) -> ExtractorResult<FastGraph> {
     match turtle::parse_bufread(buf_read).collect_triples() {
         Ok(it) => Ok(it),
@@ -38,7 +42,11 @@ pub fn load_graph_str(input_str: &str) -> ExtractorResult<FastGraph> {
 pub fn parse_str(input_str: &str) -> ExtractorResult<Document> {
     let graph = load_graph_str(input_str)?;
     let triples_maps = extract_triples_maps(&graph)?;
-    Ok(Document { triples_maps })
+    let base_iri = input_str.split('\n').filter_map(extract_base_iri).next();
+    Ok(Document {
+        triples_maps,
+        default_base_iri: base_iri,
+    })
 }
 
 pub fn parse_file(path: PathBuf) -> ExtractorResult<Document> {
@@ -50,9 +58,19 @@ pub fn parse_file(path: PathBuf) -> ExtractorResult<Document> {
             )));
         }
 
-        let buf_read = BufReader::new(File::open(path)?);
+        let buf_read = BufReader::new(File::open(path.clone())?);
         let triples_maps = extract_triples_maps(&load_graph_bread(buf_read)?)?;
-        return Ok(Document { triples_maps });
+
+        // TODO: Refactor extraction of base iri from RML file <02-08-24, SMO> //
+        let mut buf_read = BufReader::new(File::open(path)?);
+        let mut input_string = String::default();
+        buf_read.read_to_string(&mut input_string)?;
+        let base_iri = input_string.split('\n').filter_map(extract_base_iri).next();
+
+        return Ok(Document {
+            triples_maps,
+            default_base_iri: base_iri,
+        });
     }
 
     Err(ParseError::IOErrorStr(format!(

rml-interpreter/src/extractors/objectmap_extractor.rs

@@ -7,7 +7,9 @@ use super::{ExtractorResult, FromVocab, TermMapExtractor};
 use crate::extractors::store::{get_object, get_objects};
 use crate::extractors::Extractor;
 use crate::rml_model::join::JoinCondition;
-use crate::rml_model::term_map::{GraphMap, ObjectMap, TermMapInfo};
+use crate::rml_model::term_map::{
+    GraphMap, ObjectMap, TermMapInfo, TermMapType,
+};
 use crate::IriString;
 
 fn extract_join_condition(
@@ -93,8 +95,23 @@ impl TermMapExtractor<ObjectMap> for ObjectMap {
 
         let mut tm_info = tm_info_res?;
         if tm_info.term_type.is_none() {
-            tm_info.term_type = Some(tm_info.term_value.kind());
+            let mut inferred_term_type = match tm_info.term_map_type {
+                TermMapType::Reference => Some(TermKind::Literal),
+                TermMapType::Template => Some(TermKind::Iri),
+                _ => None,
+            };
+
+            if inferred_term_type.is_none() {
+                if language.is_some() || data_type.is_some() {
+                    inferred_term_type = Some(TermKind::Literal);
+                } else {
+                    inferred_term_type = Some(TermKind::Iri);
+                }
+            }
+
+            tm_info.term_type = inferred_term_type;
         }
+
         let graph_maps =
             GraphMap::extract_many_from_container(graph_ref, subj_ref)?;
 

rml-interpreter/src/extractors/term_map_info_extractor.rs

@@ -44,7 +44,10 @@ fn extract_term_map_type_value(
                     ));
     }
 
-    let trip = results_query.pop().ok_or(ParseError::GenericError("Term map doesn't have rr:constant, rr:template, rr:reference, fnml:functionValue nor rr:column.".to_string()))?;
+    let trip = results_query
+        .pop()
+        .ok_or(ParseError::GenericError("Term map doesn't have rr:constant, rr:template, rr:reference, fnml:functionValue nor rr:column.".to_string()))?;
+
     let fetched_pred = trip.p();
 
     let term_map_type_res = match fetched_pred {
@@ -79,6 +82,7 @@ impl Extractor<TermMapInfo> for TermMapInfo {
 
         let mut term_type = None;
 
+        //Explicit term type casting trough rr:termtype predicate
         if let Ok(term_type_soph) =
             get_object(graph_ref, subj_ref, &term_type_pred)
         {
@@ -100,6 +104,16 @@ impl Extractor<TermMapInfo> for TermMapInfo {
             };
         }
 
+        //Implicit term type derivation for constant-valued term maps
+        if term_map_type == TermMapType::Constant {
+            term_type = match term_value {
+                sophia_term::Term::Iri(_) => Some(TermKind::Iri),
+                sophia_term::Term::BNode(_) => Some(TermKind::BlankNode),
+                sophia_term::Term::Literal(_) => Some(TermKind::Literal),
+                sophia_term::Term::Variable(_) => None,
+            };
+        }
+
         let logical_target_iris = get_objects(
             graph_ref,
             subj_ref,

rml-interpreter/src/rml_model/mod.rs

@@ -7,7 +7,8 @@ pub mod term_map;
 
 #[derive(Debug, Clone)]
 pub struct Document {
-    pub triples_maps: Vec<TriplesMap>,
+    pub default_base_iri: Option<String>,
+    pub triples_maps:     Vec<TriplesMap>,
 }
 
 #[derive(Debug, Clone)]

rust-toolchain.toml

@@ -0,0 +1,2 @@
+[toolchain]
+channel="1.77.2"

shexml-interpreter/src/parser/tests.rs

@@ -1,6 +1,9 @@
 #[cfg(test)]
 use std::collections::HashSet;
 
+use chumsky::prelude::*; 
+use crate::*;
+
 fn assert_parse_expected<T: std::fmt::Debug + PartialEq + Eq>(
     parsed_items: Option<T>,
     expected_items: Option<T>,

src/java/test.sh

@@ -1,3 +1,3 @@
 #!/bin/sh
 
-cat ../../resources/csv-testcases/RMLTC0001a-CSV/mapping.ttl | java -Djava.library.path=$(pwd)/../../target/debug Translator
+cat ../../resources/csv-testcases/RMLTC0001a-CSV/mapping.ttl | java -Djava.library.path=$(pwd)/../../target/release Translator

src/lib.rs

@@ -4,8 +4,15 @@ pub mod shexml;
 pub mod logger;
 pub mod api;
 pub mod util;
+
+#[cfg(feature = "jni")]
 mod java;
-mod nodejs;
-mod python;
 
+#[cfg(feature = "jni")]
 pub use java::*;
+
+#[cfg(feature = "neon")]
+mod nodejs;
+
+#[cfg(feature = "pyo3")]
+mod python;