Roche::ingress_egress raises linmin error

[StuartLittlefair]

When calling ingress_egress with q=0.0578808 and i=85.379 to see if a point at 2.47587 -2.5694 0 is eclipsed by the donor star I get the following error output

Roche::linmin failed to bracket minimum, error 2

It's unclear how to deal with this.

[trmrsh]

Hi Stu,

yes, I am aware that it sometimes fails. Is this a problem for you for a
real computation or just with the py2 --> 3 issue?

Its useful to have the exact numbers here, and I will try to take a look at
it to see if I can fix it.

tom

On 12 October 2015 at 14:17, StuartLittlefair notifications@github.com
wrote:

When calling ingress_egress with q=0.0578808 and i=85.379 to see if a
point at 2.47587 -2.5694 0 is eclipsed by the donor star I get the
following error output

Roche::linmin failed to bracket minimum, error 2

It's unclear how to deal with this.

—
Reply to this email directly or view it on GitHub
#2.

[StuartLittlefair]

It's not a serious problem, but I've moved LFIT to use ingress_egress in an
attempt to speed things up. Previously I used blink a lot. It now fails
occasionally.

I'm just assuming that when it does this the point is not eclipsed for now
and swallowing the error...

On 13 October 2015 at 10:35, Tom notifications@github.com wrote:

Hi Stu,

yes, I am aware that it sometimes fails. Is this a problem for you for a
real computation or just with the py2 --> 3 issue?

Its useful to have the exact numbers here, and I will try to take a look at
it to see if I can fix it.

tom

On 12 October 2015 at 14:17, StuartLittlefair notifications@github.com
wrote:

When calling ingress_egress with q=0.0578808 and i=85.379 to see if a
point at 2.47587 -2.5694 0 is eclipsed by the donor star I get the
following error output

Roche::linmin failed to bracket minimum, error 2

It's unclear how to deal with this.

—
Reply to this email directly or view it on GitHub
#2.

—
Reply to this email directly or view it on GitHub
#2 (comment).

Stuart Littlefair

---

Dept. of Physics & Astronomy,
Univ. of Sheffield, Sheffield, S3 7RH.

email: S.Littlefair@sheffield.ac.uk
phone: +44 114 2224525

[trmrsh]

I run into it as well now and then but have always managed to escape it,
but I should see if I can fix it.

tom

On 13 October 2015 at 10:46, StuartLittlefair notifications@github.com
wrote:

It's not a serious problem, but I've moved LFIT to use ingress_egress in an
attempt to speed things up. Previously I used blink a lot. It now fails
occasionally.

I'm just assuming that when it does this the point is not eclipsed for now
and swallowing the error...

On 13 October 2015 at 10:35, Tom notifications@github.com wrote:

Hi Stu,

yes, I am aware that it sometimes fails. Is this a problem for you for a
real computation or just with the py2 --> 3 issue?

Its useful to have the exact numbers here, and I will try to take a look
at
it to see if I can fix it.

tom

On 12 October 2015 at 14:17, StuartLittlefair notifications@github.com
wrote:

When calling ingress_egress with q=0.0578808 and i=85.379 to see if a
point at 2.47587 -2.5694 0 is eclipsed by the donor star I get the
following error output

Roche::linmin failed to bracket minimum, error 2

It's unclear how to deal with this.

—
Reply to this email directly or view it on GitHub
#2.

—
Reply to this email directly or view it on GitHub
#2 (comment).

Stuart Littlefair

---

Dept. of Physics & Astronomy,
Univ. of Sheffield, Sheffield, S3 7RH.

email: S.Littlefair@sheffield.ac.uk
phone: +44 114 2224525

—
Reply to this email directly or view it on GitHub
#2 (comment).

[wildjames]

I think I'm experiencing a version of this issue. I get an error many times over an MCMC chain (the frequency seems to vary) that kicks back with this:

Diagnostics
xb,xc,fb,fc= 1,1,,-2.09253,-2.09253
q,ci,si,p = 0.0503799, 0.0777121, 0.996976, 2.41302 -2.59534 0
phi, lam, dphi, dlam = 0.672343, 2.94606, 0, -0.0312574
phi1,phi2,lam1,lam2 = 0.668958, 0.672343, 2.9148, 2.97746
pref, acc = -2.20445, 0.000177465
0 0 -2.09122 -0.0028331
1 0.02 -2.09127 -0.00277074
2 0.04 -2.09133 -0.00270842
3 0.06 -2.09138 -0.00264612
4 0.08 -2.09144 -0.00258387
5 0.1 -2.09149 -0.00252166
6 0.12 -2.09154 -0.0024595
7 0.14 -2.09158 -0.00239739
8 0.16 -2.09163 -0.00233534
9 0.18 -2.09168 -0.00227335
10 0.2 -2.09172 -0.00221143
11 0.22 -2.09177 -0.00214959
12 0.24 -2.09181 -0.00208782
13 0.26 -2.09185 -0.00202613
14 0.28 -2.09189 -0.00196453
15 0.3 -2.09193 -0.00190303
16 0.32 -2.09197 -0.00184162
17 0.34 -2.092 -0.00178031
18 0.36 -2.09204 -0.00171911
19 0.38 -2.09207 -0.00165802
20 0.4 -2.0921 -0.00159704
21 0.42 -2.09213 -0.00153619
22 0.44 -2.09216 -0.00147546
23 0.46 -2.09219 -0.00141486
24 0.48 -2.09222 -0.0013544
25 0.5 -2.09225 -0.00129407
26 0.52 -2.09227 -0.00123389
27 0.54 -2.0923 -0.00117386
28 0.56 -2.09232 -0.00111399
29 0.58 -2.09234 -0.00105427
30 0.6 -2.09236 -0.000994709
31 0.62 -2.09238 -0.00093532
32 0.64 -2.0924 -0.000876101
33 0.66 -2.09242 -0.000817059
34 0.68 -2.09243 -0.000758197
35 0.7 -2.09245 -0.00069952
36 0.72 -2.09246 -0.000641032
37 0.74 -2.09247 -0.000582738
38 0.76 -2.09248 -0.000524642
39 0.78 -2.09249 -0.000466748
40 0.8 -2.0925 -0.000409061
41 0.82 -2.09251 -0.000351584
42 0.84 -2.09252 -0.000294323
43 0.86 -2.09252 -0.000237281
44 0.88 -2.09253 -0.000180462
45 0.9 -2.09253 -0.00012387
46 0.92 -2.09253 -6.75097e-05
47 0.94 -2.09253 -1.1385e-05
48 0.96 -2.09253 4.45002e-05
49 0.98 -2.09253 0.000100142
50 1 -2.09253 0.000155536
ValueError('setup grid failed with linmin error')
Error: setup grid failed with linmin error; parlist: [0.09021958287754558, 0.026674351652483153, 0.013744000187760196, 0.00025009950737006327, 0.05037991581497417, 0.04429157722643255, 0.4095789506968387, 0.44649996165818984, 0.017277431457750726, 0.013617261084380886, 127.87002855562842, 0.43196900033098856, 1.7821816566694784, -5.177135543962317e-05, 2.521564860791719, 0.34902572616817606, 74.43299412020373, -25.668439873507573]

I can provide a decent number of parameter lists that produce this if you like, but for now I'll just post two. Here's another message:

Diagnostics
xb,xc,fb,fc= 1,1,,-2.11025,-2.11025
q,ci,si,p = 0.266106, 0.221297, 0.975206, 1.02412 1.65906 0
phi, lam, dphi, dlam = 0.250271, 1.61798, 0, -0.0206471
phi1,phi2,lam1,lam2 = 0.250271, 0.254357, 1.59733, 1.63887
pref, acc = -2.38773, 0.000144493
0 0 -2.10944 -0.0016429
1 0.02 -2.10947 -0.00160972
2 0.04 -2.1095 -0.00157654
3 0.06 -2.10953 -0.00154336
4 0.08 -2.10956 -0.00151018
5 0.1 -2.10959 -0.001477
6 0.12 -2.10962 -0.00144382
7 0.14 -2.10965 -0.00141065
8 0.16 -2.10968 -0.00137748
9 0.18 -2.10971 -0.00134431
10 0.2 -2.10973 -0.00131115
11 0.22 -2.10976 -0.00127799
12 0.24 -2.10978 -0.00124484
13 0.26 -2.10981 -0.00121169
14 0.28 -2.10983 -0.00117854
15 0.3 -2.10985 -0.0011454
16 0.32 -2.10988 -0.00111227
17 0.34 -2.1099 -0.00107915
18 0.36 -2.10992 -0.00104603
19 0.38 -2.10994 -0.00101292
20 0.4 -2.10996 -0.000979813
21 0.42 -2.10998 -0.000946718
22 0.44 -2.11 -0.000913632
23 0.46 -2.11002 -0.000880555
24 0.48 -2.11003 -0.000847487
25 0.5 -2.11005 -0.00081443
26 0.52 -2.11007 -0.000781383
27 0.54 -2.11008 -0.000748347
28 0.56 -2.1101 -0.000715322
29 0.58 -2.11011 -0.000682308
30 0.6 -2.11012 -0.000649307
31 0.62 -2.11014 -0.000616318
32 0.64 -2.11015 -0.000583341
33 0.66 -2.11016 -0.000550378
34 0.68 -2.11017 -0.000517429
35 0.7 -2.11018 -0.000484494
36 0.72 -2.11019 -0.000451573
37 0.74 -2.1102 -0.000418667
38 0.76 -2.11021 -0.000385776
39 0.78 -2.11021 -0.000352902
40 0.8 -2.11022 -0.000320043
41 0.82 -2.11023 -0.0002872
42 0.84 -2.11023 -0.000254375
43 0.86 -2.11024 -0.000221567
44 0.88 -2.11024 -0.000188776
45 0.9 -2.11024 -0.000156004
46 0.92 -2.11025 -0.00012325
47 0.94 -2.11025 -9.05157e-05
48 0.96 -2.11025 -5.78002e-05
49 0.98 -2.11025 -2.51044e-05
50 1 -2.11025 7.57122e-06
ValueError('setup grid failed with linmin error')
Error: setup grid failed with linmin error; parlist: [0.12006416024839853, 0.05492891216742712, 0.15431241965402165, 0.008651142508537202, 0.26610633216414287, 0.04194458923390031, 0.44616471570591865, 0.28385990968989805, 0.03525109079315824, 0.10064684725508243, 63.213504141460774, 0.01928129361370185, 0.08466085793302407, -0.0024301974175894026, 0.8051254579358218, 1.4497382465152464, 51.82644021750187, 65.35239013787086]

[trmrsh]

Thanks James. The job its trying to do is locate a minimum in the Roche potential between two points, and presumably not finding one. It is a long running issue that needs some work; I won't have the time for some time to tackle it I'm afraid, just got too much on. I am sure it's fixable.