Analyzing passages from a CEX source file

read_ctsdata() (in grammatike/ctsdata.py) reads a list of citable passages — each one a CTS URN paired with its own text — out of a pipe-delimited file, the input-side counterpart to write_analyses()/read_analyses() above (which handle an analysis’s results, not the passages you’re about to analyze):

from grammatike import read_ctsdata

rows = read_ctsdata("passages.txt")
for row in rows:
    citation = row.urnbase + row.citation  # reconstructs the full URN
    print(citation, "--", row.text)

The file has one or more #!ctsdata blocks, each with its own urn|text header row:

#!ctsdata
urn|text
urn:cts:greekLit:tlg0059.tlg002.perseus-grc2:17a|ὅτι μὲν ὑμεῖς, ὦ ἄνδρες Ἀθηναῖοι, πεπόνθατε ὑπὸ τῶν ἐμῶν κατηγόρων, οὐκ οἶδα·

Each row’s urn column must be a 5-part, colon-separated CTS URN; read_ctsdata() splits it into urnbase (the first 4 parts, rejoined with :, plus a trailing :) and citation (the 5th part). Pass delimiter=... if the file itself uses something other than |. Like read_analyses(), this is deliberately strict (a malformed row or a urn that doesn’t split into exactly 5 parts raises ValueError, naming the line) and merges multiple #!ctsdata blocks in file order.