Try WNYC-ANYK-1992-07-10-19332.6 Anton Batagov.WAV

E.g. http://id.loc.gov/authorities/subjects/sh92003596 (Music--Portugal)
madsrdf:Geographic is Portugal

E.g. asset info is repeated

Not working

First, do not copy entire LoC entry
Copy collection data downstream

Why

WNYC-SCHK-2004-06-24-38788.4_TK1 They.WAV

but

WNYC-SCHK-2005-10-07-43930.4_TK1 Indie Rock Tunes.WAV

??

• By default, the must be http://id.loc.gov/authorities/subjects/ or http://id.loc.gov/authorities/names/ ; otherwise error
• Make them all http (should be the URI/permalink, not the protocol)

Can be more efficient

From template "determineGeneration" in parseDAVIDTitle.xsl :

        <!-- Determine generation based on text flags such as WEB EDIT, etc 
        or from the instantiation suffix (e.g. 3b means part 2) -->
        <!-- We conisder the following Generations with two parts A:B
            A. An original master (Master:) or Derivative (Copy:)
            B. Complete (preservation or access, respectively) 
               or partial (segment)
        This means there are four possible generations:
           Master: preservation
           Copy: access
           Master: segment
           Copy: segment]
        -->

To warn of assets that should be joined

I.e., add a preferred field so that if that field is present, it is chosen over others.

If a description has a non-default value, it takes precedence

May be partly responsible for RAM issues

Create a physical manifest where you can update data for cavafy.

• Create special fake rdf with xmp extension
• Make everything updated from spreadsheet: i.e. a new cavafy ready to upload

The search accepts either input from instantiationIDs or directly in seriesQC (via a param)

For example,

https://cavafy.wnyc.org/?facet_Series+Title%5B%5D=Bob+Sherman+Production+Reels&q=%22Bob+Sherman+Production+Reels%22+%221%2F4+inch+audio+tape&search_fields%5B%5D=title&search_fields%5B%5D=format

From http://bwfmetaedit.sourceforge.net/validation_rules_help.html :

ICRD: The ICRD field must be in YYYY, YYYY-MM, YYYY-MM-DD format or initiate with a YYYY-MM-DD formatted string and then include additional free text.

Either as a separate xml or as a global parameter

Also, delete suspicious characters (\W) from filenames.

Try oXygen "generate schema" feature

I'm wonder if there is an easy way to build in “rules.” That’s to say, a reference at the top of each column to state which values in each are a passing (vs. failing) value. Like: Abstract/1, assetDate/1+, etc. Right now genre might be genre/1 but I know that you are considering changing that to genre/1+

Use these guidelines:
https://www.loc.gov/aba/publications/FreeLCGFT/J110.pdf

Also
https://www.loc.gov/aba/publications/FreeLCGFT/2020%20LCGFT%20intro.pdf

List for archival materials includes many radio genres:
https://www.loc.gov/aba/publications/FreeLCGFT/GENRE.pdf

Include a local file to choose outputs for FADGI, DAVID, etc.

Choose which reports to generate

E.g. https://id.loc.gov/authorities/subjects/sh2010102845.html

Check Copyright field in entry https://cavafy.wnyc.org/assets/1d8d983a-58af-4b75-88d5-1339a2a1cc43

Include a stoplist of LOC URIs so that the script does not climb up.

need to create warnings on the exiftool output, possibly via an attribute.

See https://xsltfiddle.liberty-development.net/jxNakAJ/57

From Statement of Work document:

Data embedded in the CodingHistory field must follow EBU Technical Recommendation R98-1999 (https://tech.ebu.ch/docs/r/r098.pdf).

Additionally, NYPR further specifically requires that the free-text (T=) parameter of the first line follows the following format:

T= [Technical information] ; [Engineer’s comments]

i.e., the technical info and engineer’s comments will be separated by a space, a semicolon, and a space.

The Technical information needed depends on the carrier format:

Analog discs

Example for a microgroove RIAA disc with fused grooves:

T=[St] 0.7 ET [RF] 50 Hz [TO] 500 Hz [RO] 13.7 dB [Sp] 33.3 rpm [Ca] 0dBFS=+18dBu ; Some fused grooves

Include the following information in this order:

Stylus size and shape, preceded by [St] and a space, expressed in thousands of an inch, followed by shape indicator.  

E. g.

[St] 2.0 ET for 2 mil elliptical truncated stylus

[St] 3.5 FCR for 3.5 mil full conical radius stylus

Compensation EQ low frequency rolloff or “rumble filter”, preceded by [RF] and a space, expressed as its corner frequency. Enter 0 if no LF rolloff or rumble filter are applied (e.g. 2-point compensation EQ). 

E.g.

[RF] 50 Hz for a 50 Hz low-frequency rolloff

Compensation EQ turnover, preceded by [TO] and a space, expressed as its corner frequency. 

E.g.

[TO] 500 Hz for NAB EQ turnover

Compensation EQ high frequency rolloff, preceded by [RO] and a space, expressed as attenuation in dB at 10000 Hz. 

E.g.

[RO] 12 dB for AES EQ rolloff (-12 dB at 10kHz)

Analog tapes

Example for a standard, 15 ips NAB tape with no noise reduction and no issues:

T=[EQ] NAB [NR] None [Sp] 15 ips [Ca] 0dBFS=+18dBu

Compensation EQ, preceded by [EQ] and a space, expressed as its common name. 

E.g.

[EQ] NAB

Noise reduction, preceded by [NR] and a space, expressed as its common name. Enter ‘None’ if no noise reduction is applied. 

E.g.

[NR] Dolby SR

Digital tapes and discs

Preemphasis, preceded by [PE] and a space, expressed as its common name (EIAJ). Enter ‘None’ if no emphasis is applied. 

E.g.

[PE] EIAJ

All formats

Playback speed, preceded by [Sp] and a space, expressed in the standard US format such as 15 ips or 33.3 rpm 

Calibration, analog sources, preceded by [Ca] and a space, expressed as dBFS=dBu to indicate the calibration of the converter in its analog-to-digital interface mode. 

E.g.

[Ca] 0dBFS=+18dBu

Calibration,  digital sources, preceded by [Ca] and a space, expressed as Slave to indicate the calibration of the converter in its digital-to-digital interface mode. 

E.g.

[Ca] Slave

Widespread transfer issues, after a space, a semicolon and a space, named following a consistent controlled vocabulary such as the AV Artifact Atlas (https://www.avartifactatlas.com/tags.html#audio). 

This will greatly reduce processing time, but it is tricky. Try building up the log and then using "for-each-group"

E.g. how it behaves depending on UUID.

Delete repeats