Fix blank node bug

DataTreehouse · Dec 16, 2023 · 63a5c70 · 63a5c70
1 parent 433f339
commit 63a5c70
Show file tree

Hide file tree

Showing 5 changed files with 13 additions and 10 deletions.
diff --git a/maplib/src/mapping.rs b/maplib/src/mapping.rs
@@ -583,6 +583,7 @@ fn create_remapped(
     let mut existing = vec![];
     let mut new = vec![];
     let mut rename_map: HashMap<&String, Vec<&String>> = HashMap::new();
+    let mut out_blank_node_counter = blank_node_counter;
 
     for (original, target) in instance
         .argument_list
@@ -624,7 +625,7 @@ fn create_remapped(
                     new_series.push(series);
                     new_dynamic_columns.insert(target_colname.clone(), primitive_column);
                     new_dynamic_from_constant.push(target_colname);
-                    blank_node_counter += input_df_height;
+                    out_blank_node_counter = max(out_blank_node_counter, blank_node_counter+input_df_height);
                 } else if original.list_expand {
                     let (expr, primitive_column) =
                         create_dynamic_expression_from_static(target_colname, ct, &target.ptype)?;
@@ -701,12 +702,13 @@ fn create_remapped(
         "Creating remapped took {} seconds",
         now.elapsed().as_secs_f32()
     );
+    println!("Remapped {}", lf.clone().collect().unwrap());
     Ok((
         lf.collect().unwrap(),
         new_dynamic_columns,
         new_constant_columns,
         new_unique_subsets,
-        blank_node_counter,
+        out_blank_node_counter,
     ))
 }
 

diff --git a/representation/Cargo.toml b/representation/Cargo.toml
@@ -10,4 +10,5 @@ oxrdf = {version = "0.1.7", features = ["oxsdatatypes"]}
 polars-core = "0.34.2"
 chrono = "0.4"
 spargebra = "0.2.8"
-bigdecimal = "0.4.1"
+bigdecimal = "0.4.1"
+polars-arrow = "0.34.2"
diff --git a/representation/src/literals.rs b/representation/src/literals.rs
@@ -1,11 +1,13 @@
 use crate::{LANG_STRING_LANG_FIELD, LANG_STRING_VALUE_FIELD};
 use chrono::{DateTime, NaiveDateTime, Utc};
 use oxrdf::vocab::rdf::LANG_STRING;
-use oxrdf::vocab::xsd;
-use oxrdf::NamedNodeRef;
+use oxrdf::vocab::{rdf, xsd};
+use oxrdf::{Literal, NamedNodeRef};
 use polars_core::datatypes::TimeUnit;
 use polars_core::prelude::{AnyValue, DataType, Field};
 use std::str::FromStr;
+use polars_arrow::offset::Offset;
+use polars_arrow::scalar::Utf8Scalar;
 
 //This code is copied and modified from Chrontext, which has identical licensing
 pub fn sparql_literal_to_any_value<'a, 'b>(
@@ -59,15 +61,13 @@ pub fn sparql_literal_to_any_value<'a, 'b>(
                 }
             }
         } else if datatype == LANG_STRING {
-            println!("VALUE: {}", value);
             let val = AnyValue::Utf8(value);
             let lang = AnyValue::Utf8(language.unwrap());
             let polars_fields: Vec<Field> = vec![
                 Field::new(LANG_STRING_VALUE_FIELD, DataType::Utf8),
                 Field::new(LANG_STRING_LANG_FIELD, DataType::Utf8),
             ];
             let av = AnyValue::StructOwned(Box::new((vec![val, lang], polars_fields)));
-            println!("AV {:?}", av);
             av
         } else {
             todo!("Not implemented! {:?}", datatype)

diff --git a/triplestore/src/lib.rs b/triplestore/src/lib.rs
@@ -1,7 +1,7 @@
 extern crate core;
 
 mod constants;
-pub(crate) mod conversion;
+pub mod conversion;
 pub mod errors;
 mod export_triples;
 mod io_funcs;

diff --git a/triplestore/src/sparql/sparql_to_polars.rs b/triplestore/src/sparql/sparql_to_polars.rs
@@ -1,5 +1,5 @@
 use chrono::NaiveDate;
-use oxrdf::vocab::xsd;
+use oxrdf::vocab::{rdf, xsd};
 use oxrdf::{BlankNode, Literal, NamedNode, Term};
 use polars::export::chrono::{DateTime, NaiveDateTime, Utc};
 use polars::prelude::{LiteralValue, NamedFrom, Series, TimeUnit};
@@ -29,7 +29,7 @@ pub fn sparql_literal_to_polars_literal_value(lit: &Literal) -> LiteralValue {
     let value = lit.value();
     let literal_value = if datatype == xsd::STRING {
         LiteralValue::Utf8(value.to_string())
-    } else if datatype == xsd::UNSIGNED_INT {
+    }  else if datatype == xsd::UNSIGNED_INT {
         let u = u32::from_str(value).expect("Integer parsing error");
         LiteralValue::UInt32(u)
     } else if datatype == xsd::UNSIGNED_LONG {