CremonaDatabase omits data for curves not first in their class

[John Cremona]

Chris Wuthrich reported that

sage: E = EllipticCurve('100467a2')
sage: E.database_attributes()
{'conductor': 100467,
 'cremona_label': '100467a2',
 'rank': 1,
 'torsion_order': 1}

although the database does contain the generators:

sage: CremonaDatabase().allgens(100467)['a2']
[[7465870518064287, 219103670535029299, 7758864174573]]

It seems that (in two palces) the code to extract the data from the database wrongly assumes that some data only exists for the first curve in each isogeny class, which is not true.

A simple fix to sage/databases/cremona.py is on its way.

[John Cremona]

New commits:

​d8f530c

17904: fix ec database data extraction when num>1

[John Cremona]

I did not add a doctest since it can only be tested with the optional database_cremona_ellcurve installed.

[wuthrich]

I agree with the fix and I checked that it solves the problem. I am running the obligatory tests now.

Thanks for fixing that so fast. Very helpful.

[Jeroen Demeyer]

Replying to cremona:

I did not add a doctest since it can only be tested with the optional database_cremona_ellcurve installed.

You should add a doctest and annotate it with

# optional - database_cremona_ellcurve

There are many such tests already in src/sage/databases/cremona.py.

(Note that I am not setting the ticket to needs_work for this)

[John Cremona]

Replying to jdemeyer:

Replying to cremona:

I did not add a doctest since it can only be tested with the optional database_cremona_ellcurve installed.

You should add a doctest and annotate it with

# optional - database_cremona_ellcurve

There are many such tests already in src/sage/databases/cremona.py.

OK, I will do that!

(Note that I am not setting the ticket to needs_work for this)

[git]

Branch pushed to git repo; I updated commit sha1 and set ticket back to needs_review. New commits:

​50a18a4

#17904 added doctests

[John Cremona]

OK, so I added the doctest and the act of pushing the new commit automatically caused the needs_review flag to be set. Sorry, Chris!

[wuthrich]

Ok. I fixed a tiny sphinx error on that page, too.

I did test cremona.py, but I have not run the complete test again after the last changes which only affect the docstring in that file. If anyone finds that inacceptable, then I will run them.

---

New commits:

​4de6008

trac 17904: small doc correction

[John Cremona]

Thanks -- I think we can leave the rest to various bots.

For what it's worth I used the cremona_curves() iterator to grab every single curve and check that len(E.gens())==E.rank(), which it did fast enough to be confident that all the data was really coming from the database.