CloudSpider: Update getSpecificRecordFromOffset() method from ChunkDataReader to skip when CHUNK_END_INDICATOR is encountered

Currently the getSpecificRecordFromOffset() method from ChunkDataReader, wrongly adds the negative value of CHUNK_END_INDICATOR to the file offset and reads the wrong set of bytes. This is issue appeared due to a related issue in OffsetRecord where the actual number of records is not stored, and during search for URL, the record search may go beyond the end of metadata file.

Also, this issue only affects the redirect report generation part of the spider (this is not accessible through UI)