Closed jillianchang closed 3 years ago
Could you check in the full set of examples? Some of the ones you mention aren't present in the PR.
PS: I think what it produces for Aen. 2.219 is correct and you just have to make the test assertion conform, then remove the unittest.skip
decorator. The same is true for Aen. 2.278.
Could you check in the full set of examples? Some of the ones you mention aren't present in the PR.
I think they are all in the PR, in the commit "added some tests" (2f4e3ef) - I think if you click the arrows on the left you can expand on the code to see all of it.
Okay, now just Aen. 2.764 (That would be, I think, an error in Pharr himself we should probably correct) plus those two lines you are currently skipping. I assume the spurious h-deletion and synizesis is no longer happening on your end?
I’m still encountering the random h-deletion and synizesis. Lines 62-132 in the file (and line 183) contain those tests- are you able to see those?
2.764 actually scans fine because our pronoun citation grammar rewrites intervocalic “j”s, since we were under the assumption that he rewrote intervocalic “i”s as “j”s. So I think he took care of that poetic license aspect by preserving the “i” there.
Can't replicate. All tests pass except 2.764.
[src]$ python scansion_test.py
...............ssF..
======================================================================
FAIL: test_aen_2_764 (__main__.ScansionTest)
----------------------------------------------------------------------
Traceback (most recent call last):
File "scansion_test.py", line 178, in test_aen_2_764
self.assertEqual(
AssertionError: '' != 'prajdadserwaːbant huːk undikwe troːia gazza'
+ prajdadserwaːbant huːk undikwe troːia gazza
----------------------------------------------------------------------
Ran 20 tests in 0.019s
FAILED (failures=1, skipped=2)
I think you may need to recompile your all.far
from scratch? It seems to be out of date somehow. (It's unlikely though not impossible that mine is out of date but also better somehow...)
To compile everything from scratch, I basically delete all far files and then do thraxmakedep all.grm
?
From the grammars
directory:
rm *.far
make -j
thraxmakedep
just generates the Makefile (the instructions for how to compile them), and in this case we don't need it at all since the Makefile is already fine.
Oh cool, so now all the tests passed for me, excluding the skipped diastole ones.
This didn't happen before, but the grammar is now applying h-deletion and synezisis even when it's not necessary. Do you know why that's happening? The overapplied synizesis lines (test_aen_1_593 and test_aen_1_682) scanned properly when I tested the lines in the individual grammar files without using synizesis.