create a standard way within corppa to combine poetry metadata and text from all reference corpora

[mnaydan]

• n/a tiny corpus
• Internet Poems
• Chadwyck-Healey

primary goal / must have:

• add tiny corpus
• method to generate merged poem metadata for use with excerpts

revise configurations / setup for reuse:

• tar.gz configs should be text_dir and allow EITHER tarball or dir path
• internet poems uses tarball file list to generate metadata (or dir path)
• other poems config metadata_path is url or local file
• update shared config file
• simplify polars schema handling (disable inference)
• raise notimplemented if tar file not handled

review/testing notes:

• able to run locally
• able to run on prosody staging vm

follow up work (later):

• fix refmatcha compilation code to use consolidated reference corpora logic #221