pyradigms 0.0.3

Creates paradigms from a table of entries with parameters.

pyradigms

Creates paradigms from a table of entries with parameters.

Getting started

pip install pyradigms

Usage

There are two separate methods; read_file creates a three-dimensional dictionary from a .csv file with a list of forms with parameters, print_paradigms creates human-readable paradigms, also in .csv format, from such a dictionary. You can either first read a file and then print the desired paradigm, or you can print a paradigm directly from a three-dimensional hash.

read_file

In order to specify what parameter should be projected onto which dimension, the arguments x, y, and z must be passed. They each take a list of strings, the strings being column names present in the .csv file. z represents the multiple paradigm tables listed vertically or in multiple files in the output. y represents the rows of a single paradigm table. x represents the columns of a single paradigm table. These dimensions must all be lists of strings, if a list contains multiple strings, those parameters will be combined in the resulting paradigm. Printed in the cells of the paradigm are the values in the column Form; if your .csv file uses a different label, specify it with read_file(target_string=<label>)

Schematic usage:

import pyradigms
pd = pyradigms.Pyradigms()
pd.read_file(
    "input.csv",
    x = ["X1", "X2"],
    y = ["Y"],
    z = ["Z"],
)
pd.print_paradigms()

Here is an illustration from the included examples: The file examples/latin_verbs.csv has the columns Form Person Number Tense Verb Mood:

Form	Person	Number	Tense	Verb	Mood
portoː	1	SG	PRS	portaːre	IND
portaːs	2	SG	PRS	portaːre	IND
portat	3	SG	PRS	portaːre	IND
portaːmus	1	PL	PRS	portaːre	IND
portaːtis	2	PL	PRS	portaːre	IND
portant	3	PL	PRS	portaːre	IND
terːeoː	1	SG	PRS	terːeːre	IND
…	…	…	…	…	…

One could for example want to combine person and number, as well as tense and mood. A separate paradigm should be produced for each verb. This is achieved as follows:

import pyradigms
pd = pyradigms.Pyradigms()
pd.read_file(
    "latin_verbs.csv",
    x = ["Person", "Number"],
    y = ["Tense", "Mood"],
    z = ["Verb"],
)
pd.print_paradigms(name="latin_verb_paradigms")

This results in the following paradigm list (in "latin_verb_paradigms.csv"):

portaːre	1SG	2SG	3SG	1PL	2PL	3PL
PRS:IND	portoː	portaːs	portat	portaːmus	portaːtis	portant
PST.IPFV:IND	portaːbam	portaːbaːs	portaːbat	portaːbaːmus	portaːbaːtis	portaːbant
FUT:IND	portaːboː	portaːbis	portaːbit	portaːbimus	portaːbitis	portaːbunt
PRS:SUBJ	portem	porteːs	portet	porteːmus	porteːtis	portent
PST.IPFV:SUBJ	portaːrem	portaːreːs	portaːret	portaːreːmus	portaːreːtis	portaːrent

terːeːre	1SG	2SG	3SG	1PL	2PL	3PL
PRS:IND	terːeoː	terːeːs	terːet	terːeːmus	terːeːtis	terːent
PST.IPFV:IND	terːeːbam	terːeːbaːs	terːeːbat	terːeːbaːmus	terːeːbaːtis	terːeːbant
FUT:IND	terːeːboː	terːeːbis	terːeːbit	terːeːbimus	terːeːbitis	terːeːbunt
PRS:SUBJ	terːream	terːeaːs	terːeat	terːeaːmus	terːeaːtis	terːeant
PST.IPFV:SUBJ	terːeːrem	terːeːres	terːeret	terːeːreːmus	terːeːreːtis	terːeːrent

petere	1SG	2SG	3SG	1PL	2PL	3PL
PRS:IND	petoː	petis	petit	petimus	petitis	petunt
PST.IPFV:IND	peteːbam	peteːbas	peteːbat	peteːbaːmus	peteːbaːtis	peteːbant
FUT:IND	petam	peteːs	petet	peteːmus	peteːtis	petent
PRS:SUBJ	petam	petaːs	petat	petaːmus	petaːtis	petant
PST.IPFV:SUBJ	peteːbar	peteːbaːris; peteːbaːre	peteːbaːtur	peteːbaːmus	peteːbaːminiː	peteːbaːtur

You can arrange and combine the parameters as you want. If you want to filter a certain parameter, you can add as many filtered_parameters as you want, and filter for a specific value. If a parameter appears on none of the three axes, and not in the filtered_parameters list, it will be ignored completely, and pyradigms will simply take the first form fulfilling all criteria! (for now)

Other options of read_file():

• The option multiple_files=True distributes the output into multiple files, which represent the z axis.
• The option display=True prints a pretty table in the command line output.
• The options x_sort_order and y_sort_order take lists which will be used to sort the output along that axis.

An example: the following code combines person and mood on the x axis and uses a very idiosyncratic sort order for that axis. Number is on the y axis, verbs on the z axis; only present tense forms are taken into account. The output is pretty printed in the terminal; the z axis is distributed across three files.

pd.read_file(
    "latin_verbs.csv",
    x = ["Person", "Mood"],
    y = ["Number"],
    z = ["Verb"],
    filtered_parameters = {"Tense": "PRS"}
)
pd.print_paradigms(
	name="example_output",
	display=True,
	single_file=False,
	x_sort_order=["1IND", "2IND", "3IND", "3SUBJ", "2SUBJ", "1SUBJ"]
)

portaːre	1IND	2IND	3IND	3SUBJ	2SUBJ	1SUBJ
SG	portoː	portaːs	portat	portet	porteːs	portem
PL	portaːmus	portaːtis	portant	portent	porteːtis	porteːmus

terːeːre	1IND	2IND	3IND	3SUBJ	2SUBJ	1SUBJ
SG	terːeoː	terːeːs	terːet	terːeat	terːeaːs	terːream
PL	terːeːmus	terːeːtis	terːent	terːeant	terːeaːtis	terːeaːmus

petere	1IND	2IND	3IND	3SUBJ	2SUBJ	1SUBJ
SG	petoː	petis	petit	petat	petaːs	petam
PL	petimus	petitis	petunt	petant	petaːtis	petaːmus

print_paradigms

If you use pyradigms in your own application, you might have a three-dimensional dictionary already ready for it, rather than constructing it with pyradigms. In that case, you can pass the argument tables to print_paradigms(); the rest works identically. If no tables argument is passed, the dictionary created by read_file() will be used. Take the following dictionary of Bernese German verb forms as an example. It has three dimensions, the first being the meaning of the verb, the second being number, and the third being person.

bernese_verbs = {"to go": {
    "SG": {"1": "kɑː",
        "2": "kɛjʃ",
        "3": "kɛjtː"},
    "PL": {"1": "kœː",
        "2":"kœːtː",
        "3":"kœː"}
}, "to say": {
    "SG": {"1": "sækə",
        "2": "sɛjʃ",
        "3": "sɛjtː"},
    "PL": {"1": "sækə",
        "2":"sækətː",
        "3":"sækə"}
    }
}

With print_paradigms(bernese_verbs), a .csv file with the following content is produced:

to go	1	2	3
SG	kɑː	kɛjʃ	kɛjtː
PL	kœː	kœːtː	kœː

to say	1	2	3
SG	sækə	sɛjʃ	sɛjtː
PL	sækə	sækətː	sækə

In the .csv file, the first layer of the three-dimensional hash is represented in the z dimension, i.e. paradigm tables stacked vertically, the second layer is represented in the y axis of the individual tables, and the third layer is represented in the x axis.