diff options
author | Henning Baldersheim <balder@yahoo-inc.com> | 2022-11-20 18:15:32 +0000 |
---|---|---|
committer | Henning Baldersheim <balder@yahoo-inc.com> | 2022-11-20 18:15:32 +0000 |
commit | 37c0e98af2ab7fbeb307019a20e572def4f1ddb6 (patch) | |
tree | 02312684c58dea559b7ade43ab30abc86c9ea0a8 | |
parent | b693d572df945b9e82798e0c40b8b7b565f6a4a8 (diff) |
Separate where you nede an actual hash, and where you want the string hashed to a double.
-rw-r--r-- | eval/src/vespa/eval/eval/basic_nodes.h | 5 | ||||
-rw-r--r-- | eval/src/vespa/eval/eval/key_gen.cpp | 6 | ||||
-rw-r--r-- | eval/src/vespa/eval/eval/llvm/llvm_wrapper.cpp | 5 | ||||
-rw-r--r-- | eval/src/vespa/eval/eval/make_tensor_function.cpp | 2 | ||||
-rw-r--r-- | eval/src/vespa/eval/eval/test/eval_spec.cpp | 10 | ||||
-rw-r--r-- | eval/src/vespa/eval/eval/test/reference_evaluation.cpp | 2 | ||||
-rw-r--r-- | searchlib/src/tests/features/prod_features.cpp | 20 | ||||
-rw-r--r-- | searchlib/src/vespa/searchlib/features/attributefeature.cpp | 6 | ||||
-rw-r--r-- | searchlib/src/vespa/searchlib/features/utils.h | 8 | ||||
-rw-r--r-- | searchlib/src/vespa/searchlib/fef/query_value.cpp | 18 | ||||
-rw-r--r-- | searchlib/src/vespa/searchlib/fef/query_value.h | 2 | ||||
-rw-r--r-- | vespalib/src/vespa/vespalib/locale/c.cpp | 2 | ||||
-rw-r--r-- | vespalib/src/vespa/vespalib/util/string_hash.cpp | 6 | ||||
-rw-r--r-- | vespalib/src/vespa/vespalib/util/string_hash.h | 4 |
14 files changed, 47 insertions, 49 deletions
diff --git a/eval/src/vespa/eval/eval/basic_nodes.h b/eval/src/vespa/eval/eval/basic_nodes.h index 2db00bbcb7d..af6ee24289a 100644 --- a/eval/src/vespa/eval/eval/basic_nodes.h +++ b/eval/src/vespa/eval/eval/basic_nodes.h @@ -6,6 +6,7 @@ #include "string_stuff.h" #include <vespa/vespalib/util/hdr_abort.h> #include <vespa/vespalib/stllike/string.h> +#include <vespa/vespalib/stllike/hash_fun.h> #include <vespa/vespalib/util/string_hash.h> #include <memory> #include <map> @@ -120,9 +121,9 @@ private: public: String(const vespalib::string &value_in) : _value(value_in) {} bool is_const_double() const override { return true; } - double get_const_double_value() const override { return hash(); } + double get_const_double_value() const override { return hash2d(_value); } const vespalib::string &value() const { return _value; } - uint32_t hash() const { return hash_code(_value.data(), _value.size()); } + uint64_t hash() const { return hashValue(_value.data(), _value.size()); } vespalib::string dump(DumpContext &) const override { return as_quoted_string(_value); } diff --git a/eval/src/vespa/eval/eval/key_gen.cpp b/eval/src/vespa/eval/eval/key_gen.cpp index a8f9700e865..6bf61f5336e 100644 --- a/eval/src/vespa/eval/eval/key_gen.cpp +++ b/eval/src/vespa/eval/eval/key_gen.cpp @@ -4,8 +4,7 @@ #include "node_visitor.h" #include "node_traverser.h" -namespace vespalib { -namespace eval { +namespace vespalib::eval { using namespace nodes; @@ -106,5 +105,4 @@ vespalib::string gen_key(const Function &function, PassParams pass_params) return key_gen.key; } -} // namespace vespalib::eval -} // namespace vespalib +} diff --git a/eval/src/vespa/eval/eval/llvm/llvm_wrapper.cpp b/eval/src/vespa/eval/eval/llvm/llvm_wrapper.cpp index 158bc91dd6a..3fafb8f8b18 100644 --- a/eval/src/vespa/eval/eval/llvm/llvm_wrapper.cpp +++ b/eval/src/vespa/eval/eval/llvm/llvm_wrapper.cpp @@ -12,7 +12,6 @@ #include <llvm/IR/IRBuilder.h> #include <llvm/IR/Intrinsics.h> #include <llvm/ExecutionEngine/ExecutionEngine.h> -#include <llvm/Analysis/Passes.h> #include <llvm/IR/DataLayout.h> #include <llvm/Transforms/Scalar.h> #include <llvm/Transforms/IPO/PassManagerBuilder.h> @@ -176,7 +175,7 @@ struct FunctionBuilder : public NodeVisitor, public NodeTraverser { params.push_back(&(*itr)); } } - ~FunctionBuilder(); + ~FunctionBuilder() override; //------------------------------------------------------------------------- @@ -385,7 +384,7 @@ struct FunctionBuilder : public NodeVisitor, public NodeTraverser { push(get_param(item.id())); } void visit(const String &item) override { - push_double(item.hash()); + push_double(item.get_const_double_value()); } void visit(const In &item) override { llvm::Value *lhs = pop_double(); diff --git a/eval/src/vespa/eval/eval/make_tensor_function.cpp b/eval/src/vespa/eval/eval/make_tensor_function.cpp index 1386196e719..72e8ac6cbde 100644 --- a/eval/src/vespa/eval/eval/make_tensor_function.cpp +++ b/eval/src/vespa/eval/eval/make_tensor_function.cpp @@ -167,7 +167,7 @@ struct TensorFunctionBuilder : public NodeVisitor, public NodeTraverser { make_inject(node, node.id()); } void visit(const String &node) override { - make_const(node, stash.create<DoubleValue>(node.hash())); + make_const(node, stash.create<DoubleValue>(node.get_const_double_value())); } void visit(const In &node) override { auto my_in = std::make_unique<In>(std::make_unique<Symbol>(0)); diff --git a/eval/src/vespa/eval/eval/test/eval_spec.cpp b/eval/src/vespa/eval/eval/test/eval_spec.cpp index 02e9c34a75e..3bd9df18438 100644 --- a/eval/src/vespa/eval/eval/test/eval_spec.cpp +++ b/eval/src/vespa/eval/eval/test/eval_spec.cpp @@ -124,11 +124,11 @@ EvalSpec::add_terminal_cases() { add_expression({}, "10").add_case({}, 10.0); add_expression({}, "100").add_case({}, 100.0); add_rule({"a", -5.0, 5.0}, "a", [](double a){ return a; }); - add_expression({}, "\"\"").add_case({}, vespalib::hash_code("")); - add_expression({}, "\"foo\"").add_case({}, vespalib::hash_code("foo")); - add_expression({}, "\"foo bar baz\"").add_case({}, vespalib::hash_code("foo bar baz")); - add_expression({}, "\">\\\\\\\"\\t\\n\\r\\f<\"").add_case({}, vespalib::hash_code(">\\\"\t\n\r\f<")); - add_expression({}, "\">\\x08\\x10\\x12\\x14<\"").add_case({}, vespalib::hash_code(">\x08\x10\x12\x14<")); + add_expression({}, "\"\"").add_case({}, vespalib::hash2d("")); + add_expression({}, "\"foo\"").add_case({}, vespalib::hash2d("foo")); + add_expression({}, "\"foo bar baz\"").add_case({}, vespalib::hash2d("foo bar baz")); + add_expression({}, "\">\\\\\\\"\\t\\n\\r\\f<\"").add_case({}, vespalib::hash2d(">\\\"\t\n\r\f<")); + add_expression({}, "\">\\x08\\x10\\x12\\x14<\"").add_case({}, vespalib::hash2d(">\x08\x10\x12\x14<")); } void diff --git a/eval/src/vespa/eval/eval/test/reference_evaluation.cpp b/eval/src/vespa/eval/eval/test/reference_evaluation.cpp index da0f5a1d1f0..acbc7463e28 100644 --- a/eval/src/vespa/eval/eval/test/reference_evaluation.cpp +++ b/eval/src/vespa/eval/eval/test/reference_evaluation.cpp @@ -145,7 +145,7 @@ struct EvalNode : public NodeVisitor { eval_param(node.id()); } void visit(const String &node) override { - eval_const(num(node.hash())); + eval_const(num(node.get_const_double_value())); } void visit(const In &node) override { auto my_op1 = [&](double a) { diff --git a/searchlib/src/tests/features/prod_features.cpp b/searchlib/src/tests/features/prod_features.cpp index 80650050b25..73b66f36172 100644 --- a/searchlib/src/tests/features/prod_features.cpp +++ b/searchlib/src/tests/features/prod_features.cpp @@ -251,13 +251,13 @@ Test::testAttribute() addScore("attribute(sebool)", 0). addScore("attribute(sfloat)", 60.5f). addScore("attribute(sdouble)", 67.5f). - addScore("attribute(sstr)", (feature_t)vespalib::hash_code("foo")). + addScore("attribute(sstr)", vespalib::hash2d("foo")). addScore("attribute(sint).count", 1). addScore("attribute(sfloat).count", 1). addScore("attribute(sstr).count", 1). addScore("attribute(udefint)", search::attribute::getUndefined<feature_t>()). addScore("attribute(udeffloat)", search::attribute::getUndefined<feature_t>()). - addScore("attribute(udefstr)", (feature_t)vespalib::hash_code("")); + addScore("attribute(udefstr)", vespalib::hash2d("")); FtFeatureTest ft(_factory, exp.getKeys()); ft.getIndexEnv().getBuilder() @@ -284,8 +284,8 @@ Test::testAttribute() addScore("attribute(aint,2)", 0). addScore("attribute(afloat,0)", 70.5f). addScore("attribute(afloat,1)", 80.5f). - addScore("attribute(astr,0)", (feature_t)vespalib::hash_code("bar")). - addScore("attribute(astr,1)", (feature_t)vespalib::hash_code("baz")). + addScore("attribute(astr,0)", vespalib::hash2d("bar")). + addScore("attribute(astr,1)", vespalib::hash2d("baz")). addScore("attribute(aint).count", 2). addScore("attribute(aint,0).count", 0). addScore("attribute(afloat).count", 2). @@ -333,10 +333,10 @@ Test::testAttribute() addScore("attribute(wsstr,foo).value", 0). addScore("attribute(wsstr,foo).weight", 0). addScore("attribute(wsstr,foo).contains", 0). - addScore("attribute(wsstr,qux).value", (feature_t)vespalib::hash_code("qux")). + addScore("attribute(wsstr,qux).value", vespalib::hash2d("qux")). addScore("attribute(wsstr,qux).weight", 11). addScore("attribute(wsstr,qux).contains", 1). - addScore("attribute(wsstr,quux).value", (feature_t)vespalib::hash_code("quux")). + addScore("attribute(wsstr,quux).value", vespalib::hash2d("quux")). addScore("attribute(wsstr,quux).weight", 12). addScore("attribute(wsstr,quux).contains", 1). addScore("attribute(wsint).count", 2). @@ -1762,10 +1762,10 @@ Test::testQuery() addScore("query(def3)", 0.0). addScore("query(val1)", 1.1). addScore("query(val2)", 2.2). - addScore("query(hash1)", vespalib::hash_code("foo")). - addScore("query(hash2)", vespalib::hash_code("2")). - addScore("query(hash3)", vespalib::hash_code("foo")). - addScore("query(hash4)", vespalib::hash_code("'foo")); + addScore("query(hash1)", vespalib::hash2d("foo")). + addScore("query(hash2)", vespalib::hash2d("2")). + addScore("query(hash3)", vespalib::hash2d("foo")). + addScore("query(hash4)", vespalib::hash2d("'foo")); FtFeatureTest ft(_factory, exp.getKeys()); ft.getIndexEnv().getProperties() .add("query(def1)", "1.0") diff --git a/searchlib/src/vespa/searchlib/features/attributefeature.cpp b/searchlib/src/vespa/searchlib/features/attributefeature.cpp index 8afb5ec1f21..dae1ccc994b 100644 --- a/searchlib/src/vespa/searchlib/features/attributefeature.cpp +++ b/searchlib/src/vespa/searchlib/features/attributefeature.cpp @@ -113,7 +113,7 @@ public: * * @param attribute The attribute vector to use. */ - SingleAttributeExecutor(const T & attribute) : _attribute(attribute) { } + explicit SingleAttributeExecutor(const T & attribute) : _attribute(attribute) { } void handle_bind_outputs(vespalib::ArrayRef<fef::NumberOrObject> outputs_in) override { fef::FeatureExecutor::handle_bind_outputs(outputs_in); auto o = outputs().get_bound(); @@ -128,7 +128,7 @@ class BoolAttributeExecutor final : public fef::FeatureExecutor { private: const SingleBoolAttribute & _attribute; public: - BoolAttributeExecutor(const SingleBoolAttribute & attribute) + explicit BoolAttributeExecutor(const SingleBoolAttribute & attribute) : _attribute(attribute) {} void execute(uint32_t docId) override { @@ -162,7 +162,7 @@ private: const attribute::IAttributeVector & _attribute; public: - CountOnlyAttributeExecutor(const attribute::IAttributeVector & attribute) : _attribute(attribute) { } + explicit CountOnlyAttributeExecutor(const attribute::IAttributeVector & attribute) : _attribute(attribute) { } void execute(uint32_t docId) override; void handle_bind_outputs(vespalib::ArrayRef<fef::NumberOrObject> outputs_in) override { fef::FeatureExecutor::handle_bind_outputs(outputs_in); diff --git a/searchlib/src/vespa/searchlib/features/utils.h b/searchlib/src/vespa/searchlib/features/utils.h index eeecd55c6da..9a8ddacb99f 100644 --- a/searchlib/src/vespa/searchlib/features/utils.h +++ b/searchlib/src/vespa/searchlib/features/utils.h @@ -57,7 +57,7 @@ inline feature_t getAsFeature(T value) */ template <> inline feature_t getAsFeature<ConstCharPtr>(ConstCharPtr value) { - return static_cast<feature_t>(vespalib::hash_code(value, strlen(value))); + return vespalib::hash2d(value, strlen(value)); } /** @@ -68,7 +68,7 @@ inline feature_t getAsFeature<ConstCharPtr>(ConstCharPtr value) { */ template <> inline feature_t getAsFeature<vespalib::stringref>(vespalib::stringref value) { - return static_cast<feature_t>(vespalib::hash_code(value)); + return vespalib::hash2d(value); } @@ -184,7 +184,7 @@ lookupTable(const search::fef::IIndexEnvironment & env, const vespalib::string & inline const search::fef::ITermFieldData * getTermFieldData(const search::fef::IQueryEnvironment &env, uint32_t termId, uint32_t fieldId) { const search::fef::ITermData *td = env.getTerm(termId); - return (td == 0) ? 0 : td->lookupField(fieldId); + return (td == nullptr) ? nullptr : td->lookupField(fieldId); } /** @@ -198,7 +198,7 @@ getTermFieldData(const search::fef::IQueryEnvironment &env, uint32_t termId, uin inline search::fef::TermFieldHandle getTermFieldHandle(const search::fef::IQueryEnvironment &env, uint32_t termId, uint32_t fieldId) { const search::fef::ITermFieldData *tfd = getTermFieldData(env, termId, fieldId); - return (tfd == 0) ? search::fef::IllegalHandle : tfd->getHandle(); + return (tfd == nullptr) ? search::fef::IllegalHandle : tfd->getHandle(); } /** diff --git a/searchlib/src/vespa/searchlib/fef/query_value.cpp b/searchlib/src/vespa/searchlib/fef/query_value.cpp index 8b881bc722f..ea17743bfc7 100644 --- a/searchlib/src/vespa/searchlib/fef/query_value.cpp +++ b/searchlib/src/vespa/searchlib/fef/query_value.cpp @@ -4,6 +4,7 @@ #include "iindexenvironment.h" #include "indexproperties.h" #include "iqueryenvironment.h" +#include <vespa/searchlib/features/utils.h> #include <vespa/document/datatype/tensor_data_type.h> #include <vespa/eval/eval/fast_value.h> #include <vespa/eval/eval/interpreted_function.h> @@ -11,7 +12,6 @@ #include <vespa/eval/eval/value_codec.h> #include <vespa/vespalib/locale/c.h> #include <vespa/vespalib/util/issue.h> -#include <vespa/vespalib/util/string_hash.h> #include <cerrno> using document::TensorDataType; @@ -59,10 +59,10 @@ as_feature(const vespalib::string& str) errno = 0; double val = vespalib::locale::c::strtod(str.c_str(), &end); if (errno != 0 || *end != '\0') { // not happy - if (str.size() > 0 && str[0] == '\'') { - val = vespalib::hash_code(str.substr(1)); + if ( ! str.empty() && str[0] == '\'') { + val = features::util::getAsFeature(vespalib::stringref(str.substr(1))); } else { - val = vespalib::hash_code(str); + val = features::util::getAsFeature(vespalib::stringref(str)); } } return val; @@ -84,7 +84,7 @@ as_tensor(const vespalib::string& expr, const ValueType& wanted_type) auto fun = Function::parse(expr); if (!fun->has_error() && (fun->num_params() == 0)) { NodeTypes types = NodeTypes(*fun, {}); - ValueType res_type = types.get_type(fun->root()); + const ValueType & res_type = types.get_type(fun->root()); if (res_type == wanted_type) { SimpleObjectParams params({}); InterpretedFunction ifun(factory, *fun, types); @@ -152,12 +152,12 @@ QueryValue::QueryValue() { } -QueryValue::QueryValue(const vespalib::string& key, const vespalib::eval::ValueType& type) +QueryValue::QueryValue(const vespalib::string& key, vespalib::eval::ValueType type) : _key(key), _name("query(" + key + ")"), _old_key("$" + key), _stored_value_key("query.value." + key), - _type(type) + _type(std::move(type)) { } @@ -171,7 +171,7 @@ QueryValue::from_config(const vespalib::string& key, const IIndexEnvironment& en if (type.is_error()) { throw InvalidValueTypeException(key, type_str); } - return {key, type}; + return {key, std::move(type)}; } std::unique_ptr<Value> @@ -187,7 +187,7 @@ QueryValue::make_default_value(const IIndexEnvironment& env) const } else { if (p.found()) { auto tensor = as_tensor(p.get(), _type); - if (tensor.get() == nullptr) { + if ( ! tensor) { throw InvalidTensorValueException(_type, p.get().c_str()); } return tensor; diff --git a/searchlib/src/vespa/searchlib/fef/query_value.h b/searchlib/src/vespa/searchlib/fef/query_value.h index fb014883268..17042d662c3 100644 --- a/searchlib/src/vespa/searchlib/fef/query_value.h +++ b/searchlib/src/vespa/searchlib/fef/query_value.h @@ -59,7 +59,7 @@ private: public: QueryValue(); - QueryValue(const vespalib::string& key, const vespalib::eval::ValueType& type); + QueryValue(const vespalib::string& key, vespalib::eval::ValueType type); ~QueryValue(); /** diff --git a/vespalib/src/vespa/vespalib/locale/c.cpp b/vespalib/src/vespa/vespalib/locale/c.cpp index f6d71857d5a..ffc29b2e104 100644 --- a/vespalib/src/vespa/vespalib/locale/c.cpp +++ b/vespalib/src/vespa/vespalib/locale/c.cpp @@ -3,7 +3,7 @@ #include "c.h" #include "locale.h" #include <cstdlib> -#include <errno.h> +#include <cerrno> namespace vespalib::locale::c { diff --git a/vespalib/src/vespa/vespalib/util/string_hash.cpp b/vespalib/src/vespa/vespalib/util/string_hash.cpp index 572ba765743..0e8b93dca04 100644 --- a/vespalib/src/vespa/vespalib/util/string_hash.cpp +++ b/vespalib/src/vespa/vespalib/util/string_hash.cpp @@ -4,7 +4,7 @@ namespace vespalib { -uint32_t hash_code(const char *str, size_t len) { +double hash2d(const char *str, size_t len) { uint32_t hash = 0; for (size_t i = 0; i < len; ++i) { hash = (hash << 5) - hash + str[i]; @@ -12,8 +12,8 @@ uint32_t hash_code(const char *str, size_t len) { return hash; } -uint32_t hash_code(vespalib::stringref str) { - return hash_code(str.data(), str.size()); +double hash2d(vespalib::stringref str) { + return hash2d(str.data(), str.size()); } } // namespace vespalib diff --git a/vespalib/src/vespa/vespalib/util/string_hash.h b/vespalib/src/vespa/vespalib/util/string_hash.h index 14a5a85a009..495a8378b42 100644 --- a/vespalib/src/vespa/vespalib/util/string_hash.h +++ b/vespalib/src/vespa/vespalib/util/string_hash.h @@ -9,8 +9,8 @@ namespace vespalib { /** * simple string hashing function similar to the one used by Java. **/ -uint32_t hash_code(const char *str, size_t len); -uint32_t hash_code(vespalib::stringref str); +double hash2d(const char *str, size_t len); +double hash2d(vespalib::stringref str); } // namespace vespalib |