Merge branch 'feature/test_large_writes' into 'master'

Github 41: Add errors for using invalid run info ids

See merge request minknow/pod5-file-format!271

CHANGELOG.md

@@ -7,6 +7,11 @@ All notable changes, updates, and fixes to pod5 will be documented here
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.2.2] 2023-06-06
+
+### Fixed
+- Add API error when adding reads with invalid end reason, pore type or run info.
+
 ## [0.2.1] 2023-05-25
 
 ### Changed

c++/pod5_format/c_api.cpp

@@ -1058,8 +1058,10 @@ inline bool load_struct_row_into_read_data(
             end_reason_internal = pod5::ReadEndReason::signal_negative;
             break;
         default:
-            pod5_set_error(
-                arrow::Status::Invalid("out of range end reason passed to pod5_add_end_reason"));
+            pod5_set_error(arrow::Status::Invalid(
+                "out of range end reason ",
+                typed_row_data->end_reason[row_id],
+                " passed to add read"));
             return false;
         }
 

c++/pod5_format/dictionary_writer.h

@@ -17,6 +17,8 @@ class POD5_FORMAT_EXPORT DictionaryWriter {
         std::shared_ptr<arrow::Array> const & indices);
     virtual pod5::Result<std::shared_ptr<arrow::Array>> get_value_array() = 0;
     virtual std::size_t item_count() = 0;
+
+    bool is_valid(std::size_t value) { return value < item_count(); }
 };
 
 }  // namespace pod5

c++/pod5_format/expandable_buffer.h

@@ -43,7 +43,10 @@ class ExpandableBuffer {
 
     gsl::span<T const> get_data_span() const
     {
-        assert(m_buffer);
+        if (!m_buffer) {
+            return {};
+        }
+
         return gsl::make_span(m_buffer->data(), m_buffer->size()).template as_span<T const>();
     }
 

c++/pod5_format/file_writer.cpp

@@ -85,6 +85,8 @@ class FileWriterImpl {
             return arrow::Status::Invalid("File writer closed, cannot write further data");
         }
 
+        ARROW_RETURN_NOT_OK(check_read(read_data));
+
         ARROW_ASSIGN_OR_RAISE(
             std::vector<std::uint64_t> signal_rows, add_signal(read_data.read_id, signal));
 
@@ -103,12 +105,31 @@ class FileWriterImpl {
             return arrow::Status::Invalid("File writer closed, cannot write further data");
         }
 
+        ARROW_RETURN_NOT_OK(check_read(read_data));
+
         // Write read data and signal row entries:
         auto read_table_row =
             m_read_table_writer->add_read(read_data, signal_rows, signal_duration);
         return read_table_row.status();
     }
 
+    arrow::Status check_read(ReadData const & read_data)
+    {
+        if (!m_read_table_dict_writers.run_info_writer->is_valid(read_data.run_info)) {
+            return arrow::Status::Invalid("Invalid run info passed to add_read");
+        }
+
+        if (!m_read_table_dict_writers.pore_writer->is_valid(read_data.pore_type)) {
+            return arrow::Status::Invalid("Invalid pore type passed to add_read");
+        }
+
+        if (!m_read_table_dict_writers.end_reason_writer->is_valid(read_data.end_reason)) {
+            return arrow::Status::Invalid("Invalid end reason passed to add_read");
+        }
+
+        return arrow::Status::OK();
+    }
+
     pod5::Result<std::vector<SignalTableRowIndex>> add_signal(
         boost::uuids::uuid const & read_id,
         gsl::span<std::int16_t const> const & signal)

c++/pod5_format/read_table_writer_utils.cpp

@@ -28,6 +28,9 @@ arrow::Result<std::shared_ptr<arrow::ArrayData>> get_array_data(
     std::size_t expected_length)
 {
     auto const value_data = builder.get_string_data();
+    if (!value_data) {
+        return Status::Invalid("Missing array value data for dictionary");
+    }
 
     arrow::TypedBufferBuilder<std::int32_t> offset_builder;
     auto const & offset_data = builder.get_typed_offset_data();