transducer: convenience option for process words iteratively (and showing output) #18118

dkrenn · 2015-04-03T08:11:41Z

This ticket introduces a new option so that iter_processs gives the output of a transducer (when processing a word) step by step. If the transducer branches (non-deterministic transducer), an exception is raised.

This is in contrast to the existing implementation, where the full output (including information on all branches of the process) is given back in each step during iteration.

An example:

sage: inverter = Transducer({'A': [('A', 0, 1), ('A', 1, 0)]},
....:     initial_states=['A'], final_states=['A'])
sage: it = inverter.iter_process(
....:     words.FibonacciWord(), iterator_type='simple')
sage: from itertools import islice
sage: [o for o in islice(it, 0, 5)]  # only output the first 5 letters
[1, 0, 1, 1, 0]

CC: @cheuberg @sagetrac-skropf

Component: finite state machines

Keywords: sd66

Author: Daniel Krenn

Branch/Commit: 9a88fc1

Reviewer: Clemens Heuberger

Issue created by migration from https://trac.sagemath.org/ticket/18118

The text was updated successfully, but these errors were encountered:

dkrenn · 2015-04-03T08:14:12Z

Branch: u/dkrenn/fsm/iter_process-simple

dkrenn · 2015-04-03T08:14:56Z

New commits:

`54c99b9`	`implement type-simple-iterator`
`9bf4d02`	`iter_process_simple: docstring and small renaming of parameter`

dkrenn · 2015-04-03T08:14:56Z

Author: Daniel Krenn

dkrenn · 2015-04-03T08:14:56Z

Commit: 9bf4d02

videlec · 2015-04-03T08:29:05Z

comment:4

Even better example

sage: inverter = Transducer({'A': [('A', 0, 1), ('A', 1, 0)]},
....:                initial_states=['A'], final_states=['A'])
sage: it = inverter.iter_process(
....: words.FibonacciWord(), iterator_type='simple')
sage: u = Words([0,1])(it)
sage: u
word: 1011010110110101101011011010110110101101...

sagetrac-git · 2015-04-03T08:36:17Z

Changed commit from 9bf4d02 to 9376b18

sagetrac-git · 2015-04-03T08:36:17Z

Branch pushed to git repo; I updated commit sha1. New commits:

`9376b18`	`improved doctests (using words)`

dkrenn · 2015-04-03T08:37:45Z

comment:6

Replying to @videlec:

Even better example

sage: inverter = Transducer({'A': [('A', 0, 1), ('A', 1, 0)]},
....:                initial_states=['A'], final_states=['A'])
sage: it = inverter.iter_process(
....: words.FibonacciWord(), iterator_type='simple')
sage: u = Words([0,1])(it)
sage: u
word: 1011010110110101101011011010110110101101...

Included. Thanks.

videlec · 2015-04-03T08:47:20Z

comment:7

Couldn't we do better? I tried to use inverter.process directly, but sadly it does not return only a word. My dream would be that the following just works

sage: W = Words([0,1])
sage: inverter.process(W([0,1,0])
word: 101
sage: inverter.process(words.FibonacciWord([0,1]))
word: 1011010110110101101011011010110110101101...

What do you think?

Vincent

dkrenn · 2015-04-03T19:23:16Z

comment:8

Replying to @videlec:

Couldn't we do better? I tried to use inverter.process directly, but sadly it does not return only a word. My dream would be that the following just works
sage: W = Words([0,1])
sage: inverter.process(W([0,1,0])
word: 101
sage: inverter.process(words.FibonacciWord([0,1]))
word: 1011010110110101101011011010110110101101...
What do you think?

Would be nice to have, but takes some effort (not sure if possible with the infinite (lazy) words). This is now #18123.

Daniel

cheuberg · 2015-04-17T08:17:54Z

Changed branch from u/dkrenn/fsm/iter_process-simple to u/cheuberg/fsm/iter_process-simple

sagetrac-git · 2015-04-17T08:19:31Z

Changed commit from 9376b18 to 137cbf9

sagetrac-git · 2015-04-17T08:19:31Z

Branch pushed to git repo; I updated commit sha1. New commits:

`a6e0fe2`	`Trac #18118: Reviewer Patch: Fix ReST formatting error`
`137cbf9`	`Trac #18118: Reviewer Patch: Add doctest`

cheuberg · 2015-04-17T08:29:04Z

comment:11

Added two minor commits, please cross-review
The description of iter_process does not explain why a user would want to have a simple iterator. Could you elaborate and/or provide an illustrating example (for simplicity, for finite input instead of an infinite word) on how the standard and the simple iterator differ?
IMHO, there should be doctests for all error conditions in _iter_process_simple_
I do not understand the code of _iter_process_simple_:
- why do you need to set output = outputs[0] and iterate over output instead of outputs[0]?
- why do you reset outputs[0] = [] at the end?
Could you please add comments on that?
Please add a SEE ALSO block for _iter_process_simple_ even if it does not currently show up in the reference manual.

cheuberg · 2015-04-17T08:29:04Z

Work Issues: see comment 10

cheuberg · 2015-04-17T08:32:34Z

Reviewer: Clemens Heuberger

dkrenn · 2015-09-09T09:24:27Z

Changed branch from u/cheuberg/fsm/iter_process-simple to u/dkrenn/fsm/iter_process-simple

dkrenn · 2015-09-09T09:25:55Z

comment:15

Replying to @cheuberg:

Added two minor commits, please cross-review

Done....ok.

The description of iter_process does not explain why a user would want to have a simple iterator. Could you elaborate and/or provide an illustrating example (for simplicity, for finite input instead of an infinite word) on how the standard and the simple iterator differ?

IMHO, there should be doctests for all error conditions in _iter_process_simple_

I do not understand the code of _iter_process_simple_:

why do you need to set output = outputs[0] and iterate over output instead of outputs[0]?

why do you reset outputs[0] = [] at the end?

Could you please add comments on that?

Please add a SEE ALSO block for _iter_process_simple_ even if it does not currently show up in the reference manual.

Addressed all these comments in the new commits.

New commits:

`4bc8bc6`	`add additional examples to show difference between 'simple'-option and not using it`
`f00e904`	`seealso block in _iter_process_simple_`
`a28221b`	`minor simplifiction of code and add a comment explaining resetting of output`
`29a54c3`	`doctests for error conditions and rephrase these errors to get more information when raised`

dkrenn · 2015-09-09T09:25:55Z

Changed commit from 137cbf9 to 29a54c3

dkrenn · 2015-09-09T09:25:55Z

Changed work issues from see comment 10 to none

cheuberg · 2015-09-13T12:53:07Z

Changed branch from u/dkrenn/fsm/iter_process-simple to u/cheuberg/fsm/iter_process-simple

cheuberg · 2015-09-13T12:58:36Z

comment:17

I added two commits, please cross-review them and set to positive_review.

New commits:

`7d9e481`	`Trac #18118: Additional doctest`
`3ecc012`	`Trac #18118: Rephrase error message.`

cheuberg · 2015-09-13T12:58:36Z

Changed commit from 29a54c3 to 3ecc012

dkrenn · 2015-09-14T07:33:17Z

Changed branch from u/cheuberg/fsm/iter_process-simple to u/dkrenn/fsm/iter_process-simple

cheuberg · 2015-09-14T14:03:49Z

comment:20

the commit field did not update, so I add a comment to do so.

New commits:

`9a88fc1`	`simplify code of a doctest (use islice)`

cheuberg · 2015-09-14T14:03:49Z

Changed commit from 3ecc012 to 9a88fc1

vbraun · 2015-09-15T07:52:27Z

Changed branch from u/dkrenn/fsm/iter_process-simple to 9a88fc1

dkrenn added this to the sage-6.6 milestone Apr 3, 2015

dkrenn added c: finite state machines labels Apr 3, 2015

dkrenn added the s: needs review label Apr 3, 2015

cheuberg added s: needs work and removed s: needs review labels Apr 17, 2015

cheuberg changed the title ~~transducer: convenience option for process words iterativly (and showing output)~~ transducer: convenience option for process words iteratively (and showing output) Apr 18, 2015

dkrenn added s: needs review and removed s: needs work labels Sep 9, 2015

cheuberg modified the milestones: sage-6.6, sage-6.9 Sep 13, 2015

dkrenn added s: positive review and removed s: needs review labels Sep 14, 2015

vbraun removed the s: positive review label Sep 15, 2015

vbraun closed this as completed in cf304f9 Sep 15, 2015

This was referenced Sep 18, 2015

automaton: iterator over words of language #15267

Closed

let Transducer(word) output a word #18123

Closed

Finite state machines: some cleanup and code rewriting #18227

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

transducer: convenience option for process words iteratively (and showing output) #18118

transducer: convenience option for process words iteratively (and showing output) #18118

dkrenn commented Apr 3, 2015

dkrenn commented Apr 3, 2015

dkrenn commented Apr 3, 2015

dkrenn commented Apr 3, 2015

dkrenn commented Apr 3, 2015

videlec commented Apr 3, 2015

sagetrac-git mannequin commented Apr 3, 2015

sagetrac-git mannequin commented Apr 3, 2015

dkrenn commented Apr 3, 2015

videlec commented Apr 3, 2015

dkrenn commented Apr 3, 2015

cheuberg commented Apr 17, 2015

sagetrac-git mannequin commented Apr 17, 2015

sagetrac-git mannequin commented Apr 17, 2015

cheuberg commented Apr 17, 2015

cheuberg commented Apr 17, 2015

cheuberg commented Apr 17, 2015

dkrenn commented Sep 9, 2015

dkrenn commented Sep 9, 2015

dkrenn commented Sep 9, 2015

dkrenn commented Sep 9, 2015

cheuberg commented Sep 13, 2015

cheuberg commented Sep 13, 2015

cheuberg commented Sep 13, 2015

dkrenn commented Sep 14, 2015

cheuberg commented Sep 14, 2015

cheuberg commented Sep 14, 2015

vbraun commented Sep 15, 2015

transducer: convenience option for process words iteratively (and showing output) #18118

transducer: convenience option for process words iteratively (and showing output) #18118

Comments

dkrenn commented Apr 3, 2015

dkrenn commented Apr 3, 2015

dkrenn commented Apr 3, 2015

dkrenn commented Apr 3, 2015

dkrenn commented Apr 3, 2015

videlec commented Apr 3, 2015

sagetrac-git mannequin commented Apr 3, 2015

sagetrac-git mannequin commented Apr 3, 2015

dkrenn commented Apr 3, 2015

videlec commented Apr 3, 2015

dkrenn commented Apr 3, 2015

cheuberg commented Apr 17, 2015

sagetrac-git mannequin commented Apr 17, 2015

sagetrac-git mannequin commented Apr 17, 2015

cheuberg commented Apr 17, 2015

cheuberg commented Apr 17, 2015

cheuberg commented Apr 17, 2015

dkrenn commented Sep 9, 2015

dkrenn commented Sep 9, 2015

dkrenn commented Sep 9, 2015

dkrenn commented Sep 9, 2015

cheuberg commented Sep 13, 2015

cheuberg commented Sep 13, 2015

cheuberg commented Sep 13, 2015

dkrenn commented Sep 14, 2015

cheuberg commented Sep 14, 2015

cheuberg commented Sep 14, 2015

vbraun commented Sep 15, 2015