tchalvak / Hidden-Sin

Hidden-Sin

This project creates a mapping of possible underlying lexical items (UR) to a surface order of words (SR/SOW), according to 13 binary parameters and a static syntactic tree. The main script is SR_creator.py, which accesses the other modules in the Modules folder.

SR_creator.py

SR_creator.py creates SRs/SOWs for all the language-agnostic URs (created by all_URs.py), for whatever languages are selected:

There is a boolean in def languages() to set to True through an argument to the script if you want all languages, otherwise all = False and it will only produce for english, ie. [0,0,0,1,0,0,1,1,0,0,0,1,1] (Or some pre-entered list of languages).

True will run all languages (ENTER EST TIME HERE)

$ python3 SR_creator.py True

False or empty arg[1] will run only English parameters (ENTER EST TIME HERE)

$ python3 SR_creator.py

all_all.txt currently looks like (No SRs are being printed currently):

0001001000011	D	S	Verb													SR:	
0001001000011	D	S	Verb													SR:	
0001001000011	D	S	Verb	Aux												SR:	
0001001000011	D	S	Verb	Adv												SR:	
0001001000011	D	S	Verb	O1												SR:	
0001001000011	D	S	Verb	PP												SR:	
0001001000011	D	S	Verb	Aux	Adv											SR:	
0001001000011	D	S	Verb	Aux	O1											SR:	
0001001000011	D	S	Verb	Aux	PP											SR:	
0001001000011	D	S	Verb	Adv	O1											SR:	
0001001000011	D	S	Verb	Adv	PP											SR:	
0001001000011	D	S	Verb	O1	O2											SR:	
0001001000011	D	S	Verb	O1	PP											SR:	
0001001000011	D	S	Verb	Aux	Adv	O1										SR:	
0001001000011	D	S	Verb	Aux	Adv	PP										SR:	
0001001000011	D	S	Verb	Aux	O1	O2										SR:	

At the end of the run, this script will run a test that compares the SR_creator output to a test.txt which contains a list of UR-SR maps known/expected to be produced by the creator.

There are two outcomes from the test:

Failure!  Test lines not found in the output!

OR

Success! Test lines found in the output!

SR Creator accesses the following modules:

nodes.py, which creates a list of node objects for each representation (i.e. a tree)
parameters.py, which applies each parameter setting according to the language that is provided
URs.py, which should be run prior to running SR_creator, as it needs to produce .txt files that SR_c will access

Folders:

grabber: this script pulls the SRs (and, optionally, the URs) form the original CoLAG data
modules: these are the modules used by SR_creator.py
S_F_Y_Data_Files: Self-explanatory
Misc: Items that got put into the Notes shared doc, etc.
EDL Learner: Original 6 parameter script, etc.
Reference Papers:
UR_writer: Scripts that create the URs for the SR_creator, these are combined in the URs.py script in the modules folder

Other:

obj_maker.py: old version of the SR_creator.py Will probably soon be fully harvested, obsolete, and deleted
all_all.txt: The output of SR_creator.py, will be created locally, as it has potential for being too large for github

Our thoughts/description/notes/analysis/interpretation of the original SFY grammar and languages can be found here