Part 4: pass inputs through algebrizing and execution. (#418)

This also adds a test that an `UnboundVariables` error is raised if a
variable mentioned in the `:in` clause isn't bound.

query-algebrizer/src/clauses/mod.rs

@@ -59,6 +59,7 @@ use types::{
     TableAlias,
 };
 
+mod inputs;
 mod or;
 mod pattern;
 mod predicate;
@@ -66,6 +67,8 @@ mod resolve;
 
 use validate::validate_or_join;
 
+pub use self::inputs::QueryInputs;
+
 // We do this a lot for errors.
 trait RcCloned<T> {
     fn cloned(&self) -> T;
@@ -240,6 +243,38 @@ impl ConjoiningClauses {
             ..Default::default()
         }
     }
+
+    pub fn with_inputs<T>(in_variables: BTreeSet<Variable>, inputs: T) -> ConjoiningClauses
+    where T: Into<Option<QueryInputs>> {
+        ConjoiningClauses::with_inputs_and_alias_counter(in_variables, inputs, RcCounter::new())
+    }
+
+    pub fn with_inputs_and_alias_counter<T>(in_variables: BTreeSet<Variable>,
+                                            inputs: T,
+                                            alias_counter: RcCounter) -> ConjoiningClauses
+    where T: Into<Option<QueryInputs>> {
+        match inputs.into() {
+            None => ConjoiningClauses::with_alias_counter(alias_counter),
+            Some(QueryInputs { mut types, mut values }) => {
+                // Discard any bindings not mentioned in our :in clause.
+                types.keep_intersected_keys(&in_variables);
+                values.keep_intersected_keys(&in_variables);
+
+                let mut cc = ConjoiningClauses {
+                    alias_counter: alias_counter,
+                    input_variables: in_variables,
+                    value_bindings: values,
+                    ..Default::default()
+                };
+
+                // Pre-fill our type mappings with the types of the input bindings.
+                cc.known_types
+                  .extend(types.iter()
+                               .map(|(k, v)| (k.clone(), unit_type_set(*v))));
+                cc
+            },
+        }
+    }
 }
 
 /// Cloning.
@@ -271,28 +306,16 @@ impl ConjoiningClauses {
     }
 }
 
-impl ConjoiningClauses {
-    #[allow(dead_code)]
-    fn with_value_bindings(bindings: BTreeMap<Variable, TypedValue>) -> ConjoiningClauses {
-        let mut cc = ConjoiningClauses {
-            value_bindings: bindings,
-            ..Default::default()
-        };
-
-        // Pre-fill our type mappings with the types of the input bindings.
-        cc.known_types
-          .extend(cc.value_bindings
-                    .iter()
-                    .map(|(k, v)| (k.clone(), unit_type_set(v.value_type()))));
-        cc
-    }
-}
-
 impl ConjoiningClauses {
     pub fn bound_value(&self, var: &Variable) -> Option<TypedValue> {
         self.value_bindings.get(var).cloned()
     }
 
+    /// Return a set of the variables externally bound to values.
+    pub fn value_bound_variables(&self) -> BTreeSet<Variable> {
+        self.value_bindings.keys().cloned().collect()
+    }
+
     /// Return a single `ValueType` if the given variable is known to have a precise type.
     /// Returns `None` if the type of the variable is unknown.
     /// Returns `None` if the type of the variable is known but not precise -- "double

query-algebrizer/src/clauses/pattern.rs

@@ -270,6 +270,7 @@ mod testing {
     use super::*;
 
     use std::collections::BTreeMap;
+    use std::collections::BTreeSet;
     use std::rc::Rc;
 
     use mentat_core::attribute::Unique;
@@ -288,6 +289,7 @@ mod testing {
     };
 
     use clauses::{
+        QueryInputs,
         add_attribute,
         associate_ident,
         ident,
@@ -660,7 +662,9 @@ mod testing {
 
         let b: BTreeMap<Variable, TypedValue> =
             vec![(y.clone(), TypedValue::Boolean(true))].into_iter().collect();
-        let mut cc = ConjoiningClauses::with_value_bindings(b);
+        let inputs = QueryInputs::with_values(b);
+        let variables: BTreeSet<Variable> = vec![Variable::from_valid_name("?y")].into_iter().collect();
+        let mut cc = ConjoiningClauses::with_inputs(variables, inputs);
 
         cc.apply_pattern(&schema, Pattern {
             source: None,
@@ -705,7 +709,9 @@ mod testing {
 
         let b: BTreeMap<Variable, TypedValue> =
             vec![(y.clone(), TypedValue::Long(42))].into_iter().collect();
-        let mut cc = ConjoiningClauses::with_value_bindings(b);
+        let inputs = QueryInputs::with_values(b);
+        let variables: BTreeSet<Variable> = vec![Variable::from_valid_name("?y")].into_iter().collect();
+        let mut cc = ConjoiningClauses::with_inputs(variables, inputs);
 
         cc.apply_pattern(&schema, Pattern {
             source: None,
@@ -737,7 +743,9 @@ mod testing {
 
         let b: BTreeMap<Variable, TypedValue> =
             vec![(y.clone(), TypedValue::Long(42))].into_iter().collect();
-        let mut cc = ConjoiningClauses::with_value_bindings(b);
+        let inputs = QueryInputs::with_values(b);
+        let variables: BTreeSet<Variable> = vec![Variable::from_valid_name("?y")].into_iter().collect();
+        let mut cc = ConjoiningClauses::with_inputs(variables, inputs);
 
         cc.apply_pattern(&schema, Pattern {
             source: None,

query-algebrizer/src/errors.rs

@@ -10,6 +10,8 @@
 
 extern crate mentat_query;
 
+use mentat_core::ValueType;
+
 use self::mentat_query::{
     PlainSymbol,
 };
@@ -20,6 +22,11 @@ error_chain! {
     }
 
     errors {
+        InputTypeDisagreement(var: PlainSymbol, declared: ValueType, provided: ValueType) {
+            description("input type disagreement")
+            display("value of type {} provided for var {}, expected {}", provided, var, declared)
+        }
+
         UnknownFunction(name: PlainSymbol) {
             description("no such function")
             display("no function named {}", name)

query-algebrizer/src/lib.rs

@@ -15,6 +15,7 @@ extern crate mentat_core;
 extern crate mentat_query;
 
 use std::collections::BTreeSet;
+use std::ops::Sub;
 
 mod errors;
 mod types;
@@ -41,6 +42,10 @@ pub use errors::{
     Result,
 };
 
+pub use clauses::{
+    QueryInputs,
+};
+
 #[allow(dead_code)]
 pub struct AlgebraicQuery {
     default_source: SrcVar,
@@ -74,16 +79,20 @@ impl AlgebraicQuery {
     pub fn is_known_empty(&self) -> bool {
         self.cc.is_known_empty()
     }
+
+    /// Return a set of the input variables mentioned in the `:in` clause that have not yet been
+    /// bound. We do this by looking at the CC.
+    pub fn unbound_variables(&self) -> BTreeSet<Variable> {
+        self.cc.input_variables.sub(&self.cc.value_bound_variables())
+    }
 }
 
 pub fn algebrize_with_counter(schema: &Schema, parsed: FindQuery, counter: usize) -> Result<AlgebraicQuery> {
-    let alias_counter = RcCounter::with_initial(counter);
+    algebrize_with_inputs(schema, parsed, counter, QueryInputs::default())
-    let cc = clauses::ConjoiningClauses::with_alias_counter(alias_counter);
-    algebrize_with_cc(schema, parsed, cc)
 }
 
 pub fn algebrize(schema: &Schema, parsed: FindQuery) -> Result<AlgebraicQuery> {
-    algebrize_with_cc(schema, parsed, clauses::ConjoiningClauses::default())
+    algebrize_with_inputs(schema, parsed, 0, QueryInputs::default())
 }
 
 /// Take an ordering list. Any variables that aren't fixed by the query are used to produce
@@ -122,8 +131,13 @@ fn validate_and_simplify_order(cc: &ConjoiningClauses, order: Option<Vec<Order>>
     }
 }
 
-#[allow(dead_code)]
+pub fn algebrize_with_inputs(schema: &Schema,
-pub fn algebrize_with_cc(schema: &Schema, parsed: FindQuery, mut cc: ConjoiningClauses) -> Result<AlgebraicQuery> {
+                             parsed: FindQuery,
+                             counter: usize,
+                             inputs: QueryInputs) -> Result<AlgebraicQuery> {
+    let alias_counter = RcCounter::with_initial(counter);
+    let mut cc = ConjoiningClauses::with_inputs_and_alias_counter(parsed.in_vars, inputs, alias_counter);
+
     // TODO: integrate default source into pattern processing.
     // TODO: flesh out the rest of find-into-context.
     let where_clauses = parsed.where_clauses;

query/src/lib.rs

@@ -34,7 +34,6 @@ extern crate edn;
 extern crate mentat_core;
 
 use std::collections::{
-    BTreeMap,
     BTreeSet,
 };
 
@@ -46,7 +45,6 @@ pub use edn::{NamespacedKeyword, PlainSymbol};
 
 use mentat_core::{
     TypedValue,
-    ValueType,
 };
 
 pub type SrcVarName = String;          // Do not include the required syntactic '$'.
@@ -743,4 +741,4 @@ impl ContainsVariables for Pattern {
             acc_ref(acc, v)
         }
     }
-}
+}

src/conn.rs

@@ -10,7 +10,6 @@
 
 #![allow(dead_code)]
 
-use std::collections::HashMap;
 use std::sync::{Arc, Mutex};
 
 use rusqlite;
@@ -19,7 +18,6 @@ use edn;
 
 use mentat_core::{
     Schema,
-    TypedValue,
 };
 
 use mentat_db::db;
@@ -29,13 +27,12 @@ use mentat_db::{
     TxReport,
 };
 
-use mentat_query::Variable;
-
 use mentat_tx_parser;
 
 use errors::*;
 use query::{
     q_once,
+    QueryInputs,
     QueryResults,
 };
 
@@ -118,7 +115,7 @@ impl Conn {
                      query: &str,
                      inputs: T,
                      limit: U) -> Result<QueryResults>
-        where T: Into<Option<HashMap<Variable, TypedValue>>>,
+        where T: Into<Option<QueryInputs>>,
               U: Into<Option<u64>>
         {
 

src/errors.rs

@@ -12,6 +12,8 @@
 
 use rusqlite;
 
+use std::collections::BTreeSet;
+
 use edn;
 use mentat_db;
 use mentat_query_algebrizer;
@@ -40,6 +42,11 @@ error_chain! {
     }
 
     errors {
+        UnboundVariables(names: BTreeSet<String>) {
+            description("unbound variables at execution time")
+            display("variables {:?} unbound at execution time", names)
+        }
+
         InvalidArgumentName(name: String) {
             description("invalid argument name")
             display("invalid argument name: '{}'", name)

src/lib.rs

@@ -54,7 +54,9 @@ pub use mentat_db::{
 pub use query::{
     NamespacedKeyword,
     PlainSymbol,
+    QueryInputs,
     QueryResults,
+    Variable,
     q_once,
 };
 

src/query.rs

@@ -8,17 +8,20 @@
 // CONDITIONS OF ANY KIND, either express or implied. See the License for the
 // specific language governing permissions and limitations under the License.
 
-use std::collections::HashMap;
-
 use rusqlite;
 use rusqlite::types::ToSql;
 
 use mentat_core::{
     Schema,
-    TypedValue,
 };
 
-use mentat_query_algebrizer::algebrize;
+use mentat_query_algebrizer::{
+    algebrize_with_inputs,
+};
+
+pub use mentat_query_algebrizer::{
+    QueryInputs,
+};
 
 pub use mentat_query::{
     NamespacedKeyword,
@@ -42,31 +45,31 @@ pub use mentat_query_projector::{
     QueryResults,
 };
 
-use errors::Result;
+use errors::{
+    ErrorKind,
+    Result,
+};
 
 pub type QueryExecutionResult = Result<QueryResults>;
 
-/// Take an EDN query string, a reference to a open SQLite connection, a Mentat DB, and an optional
+/// Take an EDN query string, a reference to an open SQLite connection, a Mentat schema, and an
-/// collection of input bindings (which should be keyed by `"?varname"`), and execute the query
+/// optional collection of input bindings (which should be keyed by `"?varname"`), and execute the
-/// immediately, blocking the current thread.
+/// query immediately, blocking the current thread.
 /// Returns a structure that corresponds to the kind of input query, populated with `TypedValue`
 /// instances.
-/// The caller is responsible for ensuring that the SQLite connection is in a transaction if
+/// The caller is responsible for ensuring that the SQLite connection has an open transaction if
 /// isolation is required.
-#[allow(unused_variables)]
 pub fn q_once<'sqlite, 'schema, 'query, T, U>
 (sqlite: &'sqlite rusqlite::Connection,
  schema: &'schema Schema,
  query: &'query str,
  inputs: T,
  limit: U) -> QueryExecutionResult
-        where T: Into<Option<HashMap<Variable, TypedValue>>>,
+        where T: Into<Option<QueryInputs>>,
               U: Into<Option<u64>>
 {
-    // TODO: validate inputs.
-
     let parsed = parse_find_string(query)?;
-    let mut algebrized = algebrize(schema, parsed)?;
+    let mut algebrized = algebrize_with_inputs(schema, parsed, 0, inputs.into().unwrap_or(QueryInputs::default()))?;
 
     if algebrized.is_known_empty() {
         // We don't need to do any SQL work at all.
@@ -75,6 +78,13 @@ pub fn q_once<'sqlite, 'schema, 'query, T, U>
 
     algebrized.apply_limit(limit.into());
 
+    // Because this is q_once, we can check that all of our `:in` variables are bound at this point.
+    // If they aren't, the user has made an error -- perhaps writing the wrong variable in `:in`, or
+    // not binding in the `QueryInput`.
+    let unbound = algebrized.unbound_variables();
+    if !unbound.is_empty() {
+        bail!(ErrorKind::UnboundVariables(unbound.into_iter().map(|v| v.to_string()).collect()));
+    }
     let select = query_to_select(algebrized);
     let SQLQuery { sql, args } = select.query.to_sql_query()?;
 

tests/query.rs

@@ -21,11 +21,18 @@ use mentat_core::{
 
 use mentat::{
     NamespacedKeyword,
+    QueryInputs,
     QueryResults,
+    Variable,
     new_connection,
     q_once,
 };
 
+use mentat::errors::{
+    Error,
+    ErrorKind,
+};
+
 #[test]
 fn test_rel() {
     let mut c = new_connection("").expect("Couldn't open conn.");
@@ -159,3 +166,41 @@ fn test_coll() {
     println!("Coll took {}µs", start.to(end).num_microseconds().unwrap());
 }
 
+#[test]
+fn test_inputs() {
+    let mut c = new_connection("").expect("Couldn't open conn.");
+    let db = mentat_db::db::ensure_current_version(&mut c).expect("Couldn't open DB.");
+
+    // entids::DB_INSTALL_VALUE_TYPE = 5.
+    let ee = (Variable::from_valid_name("?e"), TypedValue::Ref(5));
+    let inputs = QueryInputs::with_value_sequence(vec![ee]);
+    let results = q_once(&c, &db.schema,
+                         "[:find ?i . :in ?e :where [?e :db/ident ?i]]", inputs, None)
+                        .expect("query to succeed");
+
+    if let QueryResults::Scalar(Some(TypedValue::Keyword(value))) = results {
+        assert_eq!(value.as_ref(), &NamespacedKeyword::new("db.install", "valueType"));
+    } else {
+        panic!("Expected scalar.");
+    }
+}
+
+/// Ensure that a query won't be run without all of its `:in` variables being bound.
+#[test]
+fn test_unbound_inputs() {
+    let mut c = new_connection("").expect("Couldn't open conn.");
+    let db = mentat_db::db::ensure_current_version(&mut c).expect("Couldn't open DB.");
+
+    // Bind the wrong var by 'mistake'.
+    let xx = (Variable::from_valid_name("?x"), TypedValue::Ref(5));
+    let inputs = QueryInputs::with_value_sequence(vec![xx]);
+    let results = q_once(&c, &db.schema,
+                         "[:find ?i . :in ?e :where [?e :db/ident ?i]]", inputs, None);
+
+    match results {
+        Result::Err(Error(ErrorKind::UnboundVariables(vars), _)) => {
+            assert_eq!(vars, vec!["?e".to_string()].into_iter().collect());
+        },
+        _ => panic!("Expected unbound variables."),
+    }
+}