[ump45nose]

Closes #514.

What

ReaderOptions::projection is no longer required for the Avro and Parquet readers. When it is not set, the reader derives the read schema from the file schema and reads all fields.

• Avro: add FromAvroSchema(), which converts an Avro schema written by Iceberg (with field-id, element-id, key-id/value-id and the map logical type for non-string keys) back into an Iceberg Schema. Sanitized field names are restored from iceberg-field-name.
• Parquet: add FromParquetSchema(), which builds the schema from the SchemaManifest. Since it runs after ApplyNameMapping, files without field IDs work when a name mapping is provided. Fields that still have no ID (or lists/maps whose element/key/value has no ID) are skipped, because they can't be read by ID anyway. A file with neither field IDs nor a name mapping still fails as before.
• ParquetReader now builds the SchemaManifest once in Open() and uses it for both the schema derivation and the projection.

Tests

• FromAvroSchemaTest: primitive types, nested struct/list/map (string and non-string keys), sanitized names, missing field ID and unsupported types.
• ParquetSchemaConversionTest: round trip through ToParquetSchema for list and large_list, and conversion with a partial name mapping.
• Reader tests for Avro and Parquet without a projection, Parquet with a name mapping, and Parquet without field IDs or name mapping.

avro_test, parquet_test, schema_test, data_test, scan_test, manifest_test, table_test and arrow_test pass locally.

This PR was written with the help of an AI coding assistant. I reviewed the changes and tests myself.

[ump45nose]

The shared-library builds failed to link because the Parquet reader in libiceberg_bundle now calls FromArrowSchema, which wasn't exported from libiceberg. fd253e1 marks it ICEBERG_EXPORT like ToArrowSchema; a local shared build now links libiceberg_bundle.so cleanly.

(AI-assisted reply.)