MINOR: Include local file paths in reader diagnostics - #3804

Rationale for this change

When a local Parquet file is invalid, the error shows a Java object ID instead of the filename. Callers cannot identify the failing input.

Related to #3354. This fixes the built-in local input implementation, not the third-party input in that report.

What changes are included in this PR?

Return the stored path from the local input's string representation, as Hadoop-backed inputs already do.

Are these changes tested?

Testing Done

On Java 17, the same program opens a real ten-byte file through the Parquet reader.

Before, at 2df8d02:

org.apache.parquet.io.LocalInputFile@6bffbc6d is not a Parquet file (length is too low: 10)

After:

invalid.parquet is not a Parquet file (length is too low: 10)

Run from either checkout:

mvn -q -pl parquet-hadoop -am -DskipTests package dependency:build-classpath -Dmdep.outputFile=target/runtime-classpath
java -cp "parquet-common/target/classes:parquet-hadoop/target/classes:$(cat parquet-hadoop/target/runtime-classpath)" target/InvalidLocalFileProbe.java invalid.parquet

import java.nio.file.Files;
import java.nio.file.Path;
import org.apache.parquet.hadoop.ParquetFileReader;
import org.apache.parquet.io.LocalInputFile;

class InvalidLocalFileProbe {
  public static void main(String[] args) throws Exception {
    Path path = Files.write(Path.of(args[0]), new byte[10]);
    try (ParquetFileReader ignored = ParquetFileReader.open(new LocalInputFile(path))) {
      throw new AssertionError("Invalid file was accepted");
    } catch (RuntimeException exception) {
      System.out.println(exception.getMessage());
    }
  }
}

Are there any user-facing changes?

Local-file errors and reader file identifiers now show the supplied path. File contents and read behavior are unchanged.