weirdo 2.1.0__py3-none-any.whl

weirdo/matrices/helix_vs_coil.txt

@@ -0,0 +1,28 @@
+#H ZHAC000103
+#D Environment-dependent residue contact energies (rows = helix, cols = coil)
+#R PMID:10706611
+#A Zhang, C. and Kim, S.H.
+#T Environment-dependent residue contact energies for proteins
+#J Proc. Natl. Acad. Sci. USA 97, 2550-2555 (2000)
+#M rows = ARNDCQEGHILKMFPSTWYV, cols = ARNDCQEGHILKMFPSTWYV
+ 0.12  1.17  0.84 0.90 -0.81  1.16 1.44  0.10  0.69 -0.81 -0.78 1.16 -0.22 -0.67  0.61  0.47  0.36 -0.72 -0.37 -0.43
+ 0.98  1.65  1.16 0.60 -0.21  1.26 1.12  1.09  1.16 -0.04 -0.09 2.37  0.47 -0.04  1.22  1.05  0.92 -0.09  0.06  0.32
+ 0.69  1.16  1.16 1.22 -0.06  1.23 1.45  0.96  0.88  0.26  0.12 1.48  0.32  0.03  1.14  0.73  0.62  0.62  0.53  0.23
+ 0.90  0.40  1.06 1.45  0.58  1.88 2.18  1.13  0.69  0.43  0.65 0.95  0.75  0.33  1.41  0.39  0.54 -0.10  0.12  0.77
+-0.83  0.10  0.40 0.12 -2.65 -0.24 0.96 -0.26 -0.26 -1.61 -1.77 0.80 -1.02 -1.47 -0.31 -0.31 -0.49 -1.30 -0.98 -1.62
+ 1.13  1.10  1.28 1.37  0.14  1.62 1.84  1.29  1.31  0.05 -0.05 1.50  0.41  0.20  1.14  0.86  0.62  0.45  0.31  0.48
+ 1.33  0.91  1.33 1.60  0.31  1.60 1.93  1.62  1.01  0.33  0.38 1.12  0.82  0.55  1.54  0.78  0.54  0.23  0.52  0.86
+-0.22  0.72  0.27 0.47 -0.95  0.42 1.39 -0.23  0.40 -0.48 -0.81 1.04 -0.62 -0.36  0.41  0.23 -0.04 -0.71  0.08 -0.35
+ 0.47  0.81  0.95 0.51 -1.56  0.90 0.89  0.86  0.20 -0.43 -0.48 1.31 -0.63 -0.41  0.56  0.40  0.28 -0.20 -0.22 -0.21
+-0.58  0.17  0.61 0.46 -1.17  0.24 0.80  0.04 -0.16 -1.64 -1.66 0.87 -0.89 -1.56 -0.27  0.02 -0.32 -1.40 -1.13 -1.36
+-0.44  0.20  0.50 0.71 -1.56  0.11 0.82  0.28 -0.15 -1.67 -1.62 0.72 -0.96 -1.55  0.02  0.19 -0.09 -1.46 -0.95 -1.32
+ 1.07  2.48  1.75 0.98  0.42  1.68 1.04  1.31  1.39  0.41  0.29 2.95  0.98  0.27  1.63  1.51  1.48  0.32  0.60  0.64
+-0.22  0.65  0.76 0.88 -0.95  0.68 1.92  0.27  0.31 -1.32 -1.04 1.02 -0.57 -1.60  0.07  0.47  0.04 -1.29 -0.85 -0.82
+-0.33 -0.06  0.42 0.42 -1.90  0.25 0.64  0.12 -0.01 -1.64 -1.50 0.58 -1.36 -1.77 -0.30  0.02  0.04 -1.41 -1.36 -1.34
+ 0.78  1.30  1.31 1.27 -0.04  1.44 1.71  0.69  0.84  0.05  0.15 1.68  0.38  0.27  1.05  1.19  0.83 -0.24  0.23  0.12
+ 0.46  1.07  1.04 0.73 -0.31  1.47 1.23  0.57  0.58 -0.11 -0.24 1.37  0.08 -0.34  0.76  0.51  0.48 -0.04  0.47  0.18
+ 0.50  0.90  0.75 0.91 -0.26  1.03 1.25  0.55  0.55 -0.20 -0.26 1.42  0.50 -0.22  0.88  0.69  0.56  0.41  0.11 -0.15
+-0.41 -0.06 -0.19 0.32 -0.79 -0.14 0.58  0.07 -0.62 -1.58 -1.16 0.18 -1.03 -1.33 -0.56  0.15 -0.19 -1.83 -0.67 -0.92
+-0.22 -0.07  0.52 0.46 -0.87  0.38 0.59  0.40 -0.17 -1.29 -1.15 0.83 -0.98 -1.16 -0.16  0.34 -0.12 -0.79 -0.77 -0.78
+-0.51  0.49  0.48 0.67 -1.40  0.66 0.63 -0.06  0.28 -1.25 -1.50 1.14 -0.93 -1.36 -0.04  0.10 -0.01 -1.11 -0.82 -1.14
+

weirdo/matrices/helix_vs_strand.txt

@@ -0,0 +1,27 @@
+#H ZHAC000102
+#D Environment-dependent residue contact energies (rows = helix, cols = strand)
+#R PMID:10706611
+#A Zhang, C. and Kim, S.H.
+#T Environment-dependent residue contact energies for proteins
+#J Proc. Natl. Acad. Sci. USA 97, 2550-2555 (2000)
+#M rows = ARNDCQEGHILKMFPSTWYV, cols = ARNDCQEGHILKMFPSTWYV
+-0.94  1.26  0.55 0.76 -1.54  1.14 1.57 -0.78  0.44 -1.59 -1.64  1.91 -0.90 -1.49  0.28  0.20 -0.04  -0.92 -0.75 -1.45
+ 0.56  1.79  2.31 0.79 -0.67  2.54 0.72  1.09  0.94 -0.01  0.01  3.68  0.89 -0.05  1.37  0.83  1.35   0.00  0.33  0.44
+ 0.59  2.21  1.82 0.77 -0.90  0.46 3.06 -0.16  0.63 -0.33  0.20  2.43  0.99  0.63  0.54  0.24  0.63   0.11 -0.19  0.23
+ 0.66  0.76  0.76 1.19 -0.21  1.66 2.22  0.29  0.57  0.59  0.79  1.13  1.41  0.49  1.70  1.03  1.19   1.85  0.18  0.86
+-1.75  0.78 -1.00 0.32 -3.64  0.48 0.87 -1.67 -0.62 -2.77 -2.32  0.19 -1.22 -2.67 -1.62 -0.83 -1.14  -0.52 -1.94 -2.35
+ 0.33  2.15  1.22 1.26  1.37  1.17 2.56  0.92  1.02  0.11  0.00  2.58  0.79 -0.26  0.53  1.19  1.11   0.21  0.39  0.15
+ 0.82  1.05  2.18 2.11  0.01  2.42 2.58  1.15  0.97  0.20  0.31  1.31  1.25  0.12  2.00  1.09  1.13   0.58  0.31  0.39
+-0.40  0.95  0.03 0.14 -1.00  0.34 0.99 -1.32  0.13 -1.40 -1.36  1.58 -0.90 -1.41  0.82 -0.27  0.21  -0.59 -1.27 -1.09
+-0.75  2.19  0.13 0.68 -1.37  1.98 1.13  0.01  1.52 -0.83 -0.58  2.26 -0.82 -1.01  0.53 -0.17  0.02 -49.00 -0.61 -0.56
+-1.99  0.25 -0.20 1.00 -2.44 -0.12 0.88 -1.54 -0.05 -2.64 -2.33  0.75 -1.85 -2.46 -1.06 -0.59 -0.65  -1.82 -1.88 -2.45
+-2.02  0.34 -0.04 0.13 -2.29  0.24 0.73 -1.27 -0.46 -2.53 -2.44  0.67 -1.80 -2.28 -1.29 -0.40 -0.34  -1.76 -1.66 -2.26
+ 0.60  3.11  2.23 1.06  0.50  1.80 1.65  0.82  1.25  0.10  0.34  3.51  0.98 -0.21  1.15  2.09  1.30  -0.14  0.28  0.13
+-1.54 -0.06 -0.63 1.76 -2.51  0.14 0.72 -1.74  0.07 -2.27 -2.22  1.27 -1.77 -1.87  0.34 -0.02 -0.21  -0.93 -1.54 -1.81
+-2.12  0.33 -0.70 0.17 -2.30 -0.59 0.26 -1.60 -0.88 -2.53 -2.44 -0.42 -1.83 -2.68 -1.40 -0.82 -0.61  -1.63 -1.83 -2.25
+ 0.63  2.43 -0.19 1.31 -1.63  1.46 1.91  0.08  1.11 -0.20  0.47  1.94 -0.34  0.15  0.57  0.00  1.15   0.06  0.26 -0.06
+-0.41  0.88  1.02 1.04 -0.21  1.27 0.94  0.04  0.75 -0.48 -0.67  2.28  0.45 -0.92  0.75  0.50  0.96   0.22 -0.19 -0.54
+-0.32  1.48  0.35 0.43 -1.44  0.38 1.36 -0.38  0.20 -1.14 -1.00  1.38 -0.35 -0.97 -0.05 -0.16  0.29  -0.53 -0.76 -0.73
+-1.85  0.45 -0.03 0.80 -1.64 -0.23 0.11 -0.95  0.67 -1.58 -2.13  0.61 -1.75 -1.59 -1.07 -0.34 -0.40  -1.29 -1.27 -1.79
+-0.88 -0.20 -0.29 0.14 -1.31  0.09 0.71 -0.56 -0.57 -1.66 -1.38  1.40 -1.60 -1.97 -0.73 -0.32 -0.37  -1.40 -0.96 -1.38
+-1.74  0.85  0.24 0.72 -2.25  0.45 0.81 -1.29 -0.24 -2.46 -2.38  0.37 -1.21 -2.16 -1.00 -0.10 -0.57  -1.34 -1.52 -2.31

weirdo/matrices/pmbec.mat

@@ -0,0 +1,21 @@
+ A C D E F G H I K L M N P Q R S T V W Y
+A 0.322860152036 0.0113750373506 -0.0156239175966 -0.00259952715456 -0.0508792185716 0.0382679273874 -0.0832539299638 -0.00196691041626 -0.0103729638696 -0.042393907322 -0.0651042403697 -0.0853704925231 0.0757409633086 -0.0483151514798 -0.0136431408498 0.038455041596 0.0520376087986 0.081101427454 -0.125564718844 -0.0747500389698
+C 0.0113750373506 0.100680270274 0.0102951033136 0.0147570340938 0.0345785831581 0.00933463557214 -0.00750101609651 0.00476007239717 -0.0459237939975 -0.0182998264075 -0.0155971113182 0.0021128481374 -0.00860770840682 -0.0309903425175 -0.0482562439545 -0.0217965163697 -0.0227322740574 -0.0154276574266 0.0412325888637 0.00600631739163
+D -0.0156239175966 0.0102951033136 0.157208255034 0.0724279735923 -0.0189545540921 -0.00870389879389 -0.0180188107498 -0.0283467966687 -0.0634240071162 -0.0279979457557 -0.0241192288182 0.0194310374127 0.042784078891 0.000437307476866 -0.0591268568576 -0.0104660502173 0.00656101264316 -0.0193560886308 0.00415097887978 -0.0191575919464
+E -0.00259952715456 0.0147570340938 0.0724279735923 0.131775168933 -0.00519060032543 -0.00547805492393 -0.0335600821273 -0.0135417817213 -0.069471604426 0.00353800457557 -0.017166710134 0.00534055417468 0.022589833552 0.0281404974641 -0.0697402405064 -0.0172364513778 -0.0054830504799 -0.00806269508269 -0.00791955104235 -0.0231187170833
+F -0.0508792185716 0.0345785831581 -0.0189545540921 -0.00519060032543 0.259179996995 -0.00445131805782 -0.00639743486807 0.0628717025094 -0.049227253611 0.0488666377736 0.0315353570161 -0.0223593028205 -0.0919732521492 -0.0930189756622 -0.0626297946351 -0.0868415233743 -0.0777292391855 -0.015794520965 0.0625957490761 0.0858189617896
+G 0.0382679273874 0.00933463557214 -0.00870389879389 -0.00547805492393 -0.00445131805782 0.122499934434 -0.025558278086 -0.0207027221208 -0.0137316756786 -0.0326424665142 -0.0264215095016 -0.00403752148352 0.0094352664965 0.00425299048772 -0.0232280105465 0.0304312733191 0.00861853592388 -0.0127217072682 -0.0246539147339 -0.0205094859119
+H -0.0832539299638 -0.00750101609651 -0.0180188107498 -0.0335600821273 -0.00639743486807 -0.025558278086 0.207657765989 -0.0888505073496 0.0761447053198 -0.0351727012494 -0.000760393877348 0.0353903619255 -0.0682087048807 -0.00886454093107 0.109052662874 0.00938179429131 -0.0234122309305 -0.0870188771708 0.0123622841944 0.0365879336865
+I -0.00196691041626 0.00476007239717 -0.0283467966687 -0.0135417817213 0.0628717025094 -0.0207027221208 -0.0888505073496 0.27773187827 -0.0381642025534 0.0886112938313 0.0551293441776 -0.0593694184462 -0.039207153398 -0.0626883806129 -0.110160438997 -0.0618078497671 -0.0339233811197 0.091300054417 0.00138488610169 -0.0230596885329
+K -0.0103729638696 -0.0459237939975 -0.0634240071162 -0.069471604426 -0.049227253611 -0.0137316756786 0.0761447053198 -0.0381642025534 0.273355694189 -0.0177282663533 -0.00817300753785 -0.0339854863013 -0.0484016323395 -0.0331603641198 0.21516548555 0.00476731287861 -0.0331318604828 -0.0400367780545 -0.0598522551401 -0.00464804635586
+L -0.042393907322 -0.0182998264075 -0.0279979457557 0.00353800457557 0.0488666377736 -0.0326424665142 -0.0351727012494 0.0886112938313 -0.0177282663533 0.162738321535 0.0750528874999 -0.0111666419731 -0.051023845781 -0.00134001844501 -0.074934598492 -0.0584956357369 -0.031311528799 0.0449678806271 -0.00754567267671 -0.0137219703366
+M -0.0651042403697 -0.0155971113182 -0.0241192288182 -0.017166710134 0.0315353570161 -0.0264215095016 -0.000760393877348 0.0551293441776 -0.00817300753785 0.0750528874999 0.156957428383 0.00753829785887 -0.091647674076 0.00190198496329 -0.0257018542091 -0.0295349216339 -0.0454820084051 -0.0120310888206 0.0210041287765 0.0126203200265
+N -0.0853704925231 0.0021128481374 0.0194310374127 0.00534055417468 -0.0223593028205 -0.00403752148352 0.0353903619255 -0.0593694184462 -0.0339854863013 -0.0111666419731 0.00753829785887 0.151487423988 -0.0106077901881 0.0413965183445 -0.0338327997913 0.0170820288313 -0.00295174153884 -0.0436807942705 0.0296409813073 -0.00205806264345
+P 0.0757409633086 -0.00860770840682 0.042784078891 0.022589833552 -0.0919732521492 0.0094352664965 -0.0682087048807 -0.039207153398 -0.0484016323395 -0.051023845781 -0.091647674076 -0.0106077901881 0.354629507834 0.0481497903134 -0.0377142358446 -0.00687173098621 0.0199181111388 0.0225294243984 -0.0525069717881 -0.0890062760945
+Q -0.0483151514798 -0.0309903425175 0.000437307476866 0.0281404974641 -0.0930189756622 0.00425299048772 -0.00886454093107 -0.0626883806129 -0.0331603641198 -0.00134001844501 0.00190198496329 0.0413965183445 0.0481497903134 0.177175171536 0.00715630304762 0.0357241930907 0.027467611659 -0.032780800211 -0.0118972341632 -0.0487465602402
+R -0.0136431408498 -0.0482562439545 -0.0591268568576 -0.0697402405064 -0.0626297946351 -0.0232280105465 0.109052662874 -0.110160438997 0.21516548555 -0.074934598492 -0.0257018542091 -0.0338327997913 -0.0377142358446 0.00715630304762 0.389022190137 0.0204288367942 -0.0408668326839 -0.0934989556047 -0.0557627605155 0.00827128508526
+S 0.038455041596 -0.0217965163697 -0.0104660502173 -0.0172364513778 -0.0868415233743 0.0304312733191 0.00938179429131 -0.0618078497671 0.00476731287861 -0.0584956357369 -0.0295349216339 0.0170820288313 -0.00687173098621 0.0357241930907 0.0204288367942 0.161573840097 0.0839261885951 -0.00816241136786 -0.0444334801409 -0.0561239385213
+T 0.0520376087986 -0.0227322740574 0.00656101264316 -0.0054830504799 -0.0777292391855 0.00861853592388 -0.0234122309305 -0.0339233811197 -0.0331318604828 -0.031311528799 -0.0454820084051 -0.00295174153884 0.0199181111388 0.027467611659 -0.0408668326839 0.0839261885951 0.142525860495 0.0493244941272 -0.0264928932645 -0.0468623824337
+V 0.081101427454 -0.0154276574266 -0.0193560886308 -0.00806269508269 -0.015794520965 -0.0127217072682 -0.0870188771708 0.091300054417 -0.0400367780545 0.0449678806271 -0.0120310888206 -0.0436807942705 0.0225294243984 -0.032780800211 -0.0934989556047 -0.00816241136786 0.0493244941272 0.172778293246 -0.0289445753682 -0.0444846240282
+W -0.125564718844 0.0412325888637 0.00415097887978 -0.00791955104235 0.0625957490761 -0.0246539147339 0.0123622841944 0.00138488610169 -0.0598522551401 -0.00754567267671 0.0210041287765 0.0296409813073 -0.0525069717881 -0.0118972341632 -0.0557627605155 -0.0444334801409 -0.0264928932645 -0.0289445753682 0.194048086876 0.0791543436022
+Y -0.0747500389698 0.00600631739163 -0.0191575919464 -0.0231187170833 0.0858189617896 -0.0205094859119 0.0365879336865 -0.0230596885329 -0.00464804635586 -0.0137219703366 0.0126203200265 -0.00205806264345 -0.0890062760945 -0.0487465602402 0.00827128508526 -0.0561239385213 -0.0468623824337 -0.0444846240282 0.0791543436022 0.237788221516

weirdo/matrices/strand_vs_coil.txt

@@ -0,0 +1,27 @@
+#H ZHAC000105
+#D Environment-dependent residue contact energies (rows = strand, cols = coil)
+#R PMID:10706611
+#A Zhang, C. and Kim, S.H.
+#T Environment-dependent residue contact energies for proteins
+#J Proc. Natl. Acad. Sci. USA 97, 2550-2555 (2000)
+#M rows = ARNDCQEGHILKMFPSTWYV, cols = ARNDCQEGHILKMFPSTWYV
+-0.57  0.47 0.30 0.62 -1.60 0.45 0.61 -0.24  0.07 -1.64 -1.63 0.62 -1.03 -1.55 -0.11 -0.10 -0.34 -1.44 -0.39 -1.55
+ 0.23  0.79 0.76 0.39 -0.41 0.92 0.76  0.52  0.51 -0.30  0.13 1.58  0.88 -0.07  0.60  0.65  0.37  0.14  0.32  0.17
+-0.28  0.74 0.57 0.87 -0.68 0.52 1.00 -0.07  0.32 -0.31 -0.08 0.87  0.29 -0.17  0.57  0.11  0.19  0.04  0.24 -0.23
+ 0.15 -0.25 0.46 0.69 -0.46 0.41 1.34  0.56 -0.51 -0.23  0.27 0.59  0.60 -0.38  1.02  0.08  0.05 -0.48  0.02  0.34
+-1.19 -0.46 0.21 0.51 -3.30 0.26 0.20 -1.03 -0.72 -1.55 -1.71 0.27 -1.24 -1.70 -0.50 -0.55 -0.97 -0.67 -1.26 -1.62
+ 0.63  1.18 0.92 1.37 -0.30 0.93 1.27  0.56  0.91 -0.28 -0.11 0.98  0.15 -0.30  0.64  0.88  0.68 -0.44  0.66  0.15
+ 0.97  0.89 1.37 1.89  0.30 1.25 2.34  0.98  0.58  0.20  0.50 0.67  1.23  0.58  1.26  0.95  1.06  0.04  0.87  0.48
+-0.64  0.12 0.27 0.31 -1.37 0.38 0.98 -0.40 -0.12 -1.58 -1.40 0.78 -0.46 -1.38 -0.21  0.05 -0.26 -1.41 -0.61 -1.13
+-0.02  0.75 0.68 0.14 -0.58 0.73 0.84  0.41 -0.64 -0.75  0.03 1.46 -0.16 -0.49  0.52  0.31 -0.11 -1.00 -0.58  0.03
+-0.94 -0.14 0.31 0.26 -1.70 0.07 0.46 -0.37 -0.50 -1.88 -1.79 0.84 -0.99 -1.82 -0.47 -0.05 -0.54 -1.65 -1.09 -1.64
+-0.76  0.32 0.43 0.25 -1.63 0.22 0.68 -0.17 -0.40 -1.84 -1.70 0.47 -1.06 -1.76 -0.39  0.09 -0.42 -1.81 -1.15 -1.64
+ 1.02  1.99 1.18 0.59  0.08 1.10 0.60  0.61  0.95  0.24  0.34 2.69  0.97 -0.03  1.23  1.07  0.83  0.00  0.26  0.36
+-0.16  0.83 0.47 0.92 -1.63 0.36 0.71 -0.20  0.90 -1.00 -1.12 1.55 -0.31 -1.35 -0.01  0.34  0.20 -1.70 -0.60 -0.79
+-0.70  0.03 0.63 0.15 -1.26 0.29 0.35 -0.11 -0.36 -1.73 -1.55 0.71 -0.97 -1.55 -0.28 -0.09 -0.32 -1.23 -0.91 -1.30
+ 0.17  0.50 0.60 0.67 -1.31 0.50 0.94  0.02 -0.45 -1.26 -0.91 1.08  0.83 -0.87  0.63  0.31  0.26 -0.50 -0.55 -0.79
+-0.06  0.99 0.73 0.86 -0.89 0.85 0.67  0.08  0.06 -0.22 -0.29 0.94 -0.08 -0.41  0.67  0.33  0.13 -1.01  0.13 -0.24
+ 0.26  0.93 0.70 0.87 -0.78 0.58 1.20  0.12  0.52 -0.30 -0.24 1.11  0.01 -0.08  0.65  0.47  0.41 -0.31  0.12 -0.32
+-0.03 -0.11 0.27 0.66 -1.50 0.65 0.50 -0.12 -0.32 -1.13 -1.01 0.52 -1.08 -1.04 -0.32 -0.03 -0.10 -0.67 -0.73 -0.64
+-0.44  0.20 0.20 0.20 -1.26 0.16 0.10 -0.21 -0.52 -1.26 -1.30 0.60 -0.76 -1.17 -0.42  0.05 -0.27 -1.20 -0.75 -0.84
+-0.83  0.20 0.48 0.62 -1.44 0.17 0.73 -0.12 -0.26 -1.64 -1.59 0.52 -0.70 -1.55 -0.28  0.12 -0.17 -1.16 -0.85 -1.42

weirdo/model_manager.py

@@ -0,0 +1,346 @@
+"""Model manager for trained foreignness scorers.
+
+Handles saving, loading, and listing trained ML models.
+"""
+
+import json
+import os
+from datetime import datetime
+from pathlib import Path
+from typing import Any, Dict, List, Optional, Union
+
+# Default model storage directory
+DEFAULT_MODEL_DIR = Path.home() / '.weirdo' / 'models'
+
+
+class ModelInfo:
+    """Information about a saved model."""
+
+    def __init__(
+        self,
+        name: str,
+        scorer_type: str,
+        path: Path,
+        created: Optional[str] = None,
+        params: Optional[Dict[str, Any]] = None,
+        metadata: Optional[Dict[str, Any]] = None,
+    ):
+        self.name = name
+        self.scorer_type = scorer_type
+        self.path = path
+        self.created = created
+        self.params = params or {}
+        self.metadata = metadata or {}
+
+    def __repr__(self) -> str:
+        return f"ModelInfo(name='{self.name}', type='{self.scorer_type}')"
+
+    def to_dict(self) -> Dict[str, Any]:
+        """Convert to dictionary."""
+        return {
+            'name': self.name,
+            'scorer_type': self.scorer_type,
+            'path': str(self.path),
+            'created': self.created,
+            'params': self.params,
+            'metadata': self.metadata,
+        }
+
+
+class ModelManager:
+    """Manager for trained foreignness scoring models.
+
+    Handles saving, loading, and listing trained models in a
+    centralized directory structure.
+
+    Parameters
+    ----------
+    model_dir : str or Path, optional
+        Directory for storing models. Defaults to ~/.weirdo/models.
+
+    Example
+    -------
+    >>> mm = ModelManager()
+    >>> mm.list_models()  # List available models
+    >>> model = mm.load('my-mlp')  # Load a saved model
+    >>> mm.save(scorer, 'my-new-model')  # Save a trained model
+    """
+
+    def __init__(self, model_dir: Optional[Union[str, Path]] = None):
+        self._model_dir = Path(model_dir) if model_dir else DEFAULT_MODEL_DIR
+        self._model_dir.mkdir(parents=True, exist_ok=True)
+
+    @property
+    def model_dir(self) -> Path:
+        """Get model storage directory."""
+        return self._model_dir
+
+    def list_models(self) -> List[ModelInfo]:
+        """List all available models.
+
+        Returns
+        -------
+        models : list of ModelInfo
+            Information about each saved model.
+        """
+        models = []
+
+        if not self._model_dir.exists():
+            return models
+
+        for path in sorted(self._model_dir.iterdir()):
+            if path.is_dir() and (path / 'config.json').exists():
+                try:
+                    info = self._load_model_info(path)
+                    models.append(info)
+                except Exception:
+                    # Skip corrupted models
+                    pass
+
+        return models
+
+    def _load_model_info(self, path: Path) -> ModelInfo:
+        """Load model info from a model directory."""
+        with open(path / 'config.json', 'r') as f:
+            config = json.load(f)
+
+        metadata = {}
+        if (path / 'metadata.json').exists():
+            with open(path / 'metadata.json', 'r') as f:
+                metadata = json.load(f)
+
+        # Get creation time
+        created = None
+        stat = path.stat()
+        created = datetime.fromtimestamp(stat.st_mtime).isoformat()
+
+        return ModelInfo(
+            name=path.name,
+            scorer_type=config.get('scorer_type', 'unknown'),
+            path=path,
+            created=created,
+            params=config.get('params', {}),
+            metadata=metadata,
+        )
+
+    def get_model_info(self, name: str) -> Optional[ModelInfo]:
+        """Get info for a specific model.
+
+        Parameters
+        ----------
+        name : str
+            Model name.
+
+        Returns
+        -------
+        info : ModelInfo or None
+            Model info if found, None otherwise.
+        """
+        path = self._model_dir / name
+        if path.exists() and (path / 'config.json').exists():
+            return self._load_model_info(path)
+        return None
+
+    def load(self, name: str) -> 'TrainableScorer':
+        """Load a trained model by name.
+
+        Parameters
+        ----------
+        name : str
+            Model name (directory name in model storage).
+
+        Returns
+        -------
+        scorer : TrainableScorer
+            Loaded model ready for inference.
+        """
+        from .scorers.trainable import TrainableScorer
+        from .scorers.mlp import MLPScorer
+
+        path = self._model_dir / name
+        if not path.exists():
+            raise FileNotFoundError(f"Model not found: {name}")
+
+        # Load config to determine scorer type
+        with open(path / 'config.json', 'r') as f:
+            config = json.load(f)
+
+        scorer_type = config.get('scorer_type', '')
+
+        # Map to scorer class
+        scorer_classes = {
+            'MLPScorer': MLPScorer,
+        }
+
+        if scorer_type not in scorer_classes:
+            raise ValueError(
+                f"Unknown scorer type: {scorer_type}. "
+                f"Available: {list(scorer_classes.keys())}"
+            )
+
+        return scorer_classes[scorer_type].load(path)
+
+    def save(
+        self,
+        scorer: 'TrainableScorer',
+        name: str,
+        overwrite: bool = False,
+    ) -> Path:
+        """Save a trained model.
+
+        Parameters
+        ----------
+        scorer : TrainableScorer
+            Trained model to save.
+        name : str
+            Name for the saved model.
+        overwrite : bool, default=False
+            Whether to overwrite existing model with same name.
+
+        Returns
+        -------
+        path : Path
+            Path where model was saved.
+        """
+        path = self._model_dir / name
+
+        if path.exists() and not overwrite:
+            raise FileExistsError(
+                f"Model already exists: {name}. "
+                "Use overwrite=True to replace."
+            )
+
+        scorer.save(path)
+        return path
+
+    def delete(self, name: str) -> bool:
+        """Delete a saved model.
+
+        Parameters
+        ----------
+        name : str
+            Model name to delete.
+
+        Returns
+        -------
+        deleted : bool
+            True if model was deleted.
+        """
+        import shutil
+        path = self._model_dir / name
+
+        if path.exists():
+            shutil.rmtree(path)
+            return True
+        return False
+
+    def print_models(self) -> None:
+        """Print formatted list of available models."""
+        models = self.list_models()
+
+        if not models:
+            print("No trained models found.")
+            print(f"Model directory: {self._model_dir}")
+            return
+
+        print(f"Trained models ({len(models)}):")
+        print("-" * 60)
+
+        for model in models:
+            print(f"\n  {model.name}")
+            print(f"    Type: {model.scorer_type}")
+            if model.created:
+                print(f"    Created: {model.created[:19]}")
+            if 'n_train' in model.metadata:
+                print(f"    Training samples: {model.metadata['n_train']}")
+            if 'n_epochs' in model.metadata:
+                print(f"    Epochs trained: {model.metadata['n_epochs']}")
+            if 'final_train_loss' in model.metadata:
+                print(f"    Final train loss: {model.metadata['final_train_loss']:.4f}")
+            if 'best_val_loss' in model.metadata:
+                print(f"    Best val loss: {model.metadata['best_val_loss']:.4f}")
+            if 'k' in model.params:
+                print(f"    K-mer size: {model.params['k']}")
+
+        print(f"\nModel directory: {self._model_dir}")
+
+
+# Singleton instance
+_model_manager: Optional[ModelManager] = None
+
+
+def get_model_manager(model_dir: Optional[Union[str, Path]] = None) -> ModelManager:
+    """Get the model manager instance.
+
+    Parameters
+    ----------
+    model_dir : str or Path, optional
+        Custom model directory.
+
+    Returns
+    -------
+    manager : ModelManager
+    """
+    global _model_manager
+    if _model_manager is None or model_dir is not None:
+        _model_manager = ModelManager(model_dir)
+    return _model_manager
+
+
+def list_models(model_dir: Optional[Union[str, Path]] = None) -> List[ModelInfo]:
+    """List all available trained models.
+
+    Parameters
+    ----------
+    model_dir : str or Path, optional
+        Custom model directory.
+
+    Returns
+    -------
+    models : list of ModelInfo
+    """
+    return get_model_manager(model_dir).list_models()
+
+
+def load_model(name: str, model_dir: Optional[Union[str, Path]] = None) -> 'TrainableScorer':
+    """Load a trained model by name.
+
+    Parameters
+    ----------
+    name : str
+        Model name.
+    model_dir : str or Path, optional
+        Custom model directory.
+
+    Returns
+    -------
+    scorer : TrainableScorer
+        Loaded model.
+    """
+    return get_model_manager(model_dir).load(name)
+
+
+def save_model(
+    scorer: 'TrainableScorer',
+    name: str,
+    model_dir: Optional[Union[str, Path]] = None,
+    overwrite: bool = False,
+) -> Path:
+    """Save a trained model.
+
+    Parameters
+    ----------
+    scorer : TrainableScorer
+        Trained model.
+    name : str
+        Model name.
+    model_dir : str or Path, optional
+        Custom model directory.
+    overwrite : bool, default=False
+        Overwrite existing model.
+
+    Returns
+    -------
+    path : Path
+    """
+    return get_model_manager(model_dir).save(scorer, name, overwrite)

weirdo/peptide_vectorizer.py

@@ -0,0 +1,78 @@
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import numpy as np
+from sklearn.feature_extraction.text import CountVectorizer
+from sklearn.preprocessing import normalize
+
+def make_count_vectorizer(reduced_alphabet, max_ngram):
+    if reduced_alphabet is None:
+        preprocessor = None
+    else:
+        preprocessor = lambda s: "".join([reduced_alphabet[si] for si in s])
+
+    return CountVectorizer(
+        analyzer='char',
+        ngram_range=(1, max_ngram),
+        dtype=np.float64,
+        preprocessor=preprocessor)
+
+class PeptideVectorizer(object):
+    """
+    Make n-gram frequency vectors from peptide sequences
+    """
+    def __init__(
+            self,
+            max_ngram=1,
+            normalize_row=True,
+            reduced_alphabet=None,
+            training_already_reduced=False):
+        self.reduced_alphabet = reduced_alphabet
+        self.max_ngram = max_ngram
+        self.normalize_row = normalize_row
+        self.training_already_reduced = training_already_reduced
+        self.count_vectorizer = None
+
+    def __getstate__(self):
+        return {
+            'reduced_alphabet': self.reduced_alphabet,
+            'count_vectorizer': self.count_vectorizer,
+            'training_already_reduced': self.training_already_reduced,
+            'normalize_row': self.normalize_row,
+            'max_ngram': self.max_ngram,
+        }
+
+    def fit_transform(self, amino_acid_strings):
+        self.count_vectorizer = \
+            make_count_vectorizer(self.reduced_alphabet, self.max_ngram)
+
+        if self.training_already_reduced:
+            c = make_count_vectorizer(None, self.max_ngram)
+            X = c.fit_transform(amino_acid_strings).toarray()
+            self.count_vectorizer.vocabulary_ = c.vocabulary_
+        else:
+            c = self.count_vectorizer
+            X = c.fit_transform(amino_acid_strings).toarray()
+
+        if self.normalize_row:
+            X = normalize(X, norm='l1')
+        return X
+
+    def fit(self, amino_acid_strings):
+        self.fit_transform(amino_acid_strings)
+
+    def transform(self, amino_acid_strings):
+        assert self.count_vectorizer, "Must call 'fit' before 'transform'"
+        X = self.count_vectorizer.transform(amino_acid_strings).toarray()
+        if self.normalize_row:
+            X = normalize(X, norm='l1')
+        return X

weirdo/pmbec.py

@@ -0,0 +1,85 @@
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from os.path import join
+
+from .static_data import MATRIX_DIR
+
+from .amino_acid_alphabet import dict_to_amino_acid_matrix
+
+def read_pmbec_coefficients(
+        key_type='row',
+        verbose=True,
+        filename=join(MATRIX_DIR, 'pmbec.mat')):
+    """
+    Parameters
+    ------------
+
+    filename : str
+        Location of PMBEC coefficient matrix
+
+    key_type : str
+        'row' : every key is a single amino acid,
+           which maps to a dictionary for that row
+        'pair' : every key is a tuple of amino acids
+        'pair_string' : every key is a string of two amino acid characters
+
+    verbose : bool
+        Print rows of matrix as we read them
+    """
+    d = {}
+    if key_type == 'row':
+        def add_pair(row_letter, col_letter, value):
+            if row_letter not in d:
+                d[row_letter] = {}
+            d[row_letter][col_letter] = value
+    elif key_type == 'pair':
+        def add_pair(row_letter, col_letter, value):
+            d[(row_letter, col_letter)] = value
+
+    else:
+        assert key_type == 'pair_string', \
+            "Invalid dictionary key type: %s" % key_type
+
+        def add_pair(row_letter, col_letter, value):
+            d["%s%s" % (row_letter, col_letter)] = value
+
+    with open(filename, 'r') as f:
+        lines = [line for line in f.read().split('\n') if len(line) > 0]
+        header = lines[0]
+        if verbose:
+            print(header)
+        residues = [
+            x for x in header.split()
+            if len(x) == 1 and x != ' ' and x != '\t'
+        ]
+        assert len(residues) == 20
+        if verbose:
+            print(residues)
+        for line in lines[1:]:
+            cols = [
+                x
+                for x in line.split(' ')
+                if len(x) > 0 and x != ' ' and x != '\t'
+            ]
+            assert len(cols) == 21, "Expected 20 values + letter, got %s" % cols
+            row_letter = cols[0]
+            for i, col in enumerate(cols[1:]):
+                col_letter = residues[i]
+                assert col_letter != ' ' and col_letter != '\t'
+                value = float(col)
+                add_pair(row_letter, col_letter, value)
+    return d
+
+# dictionary of PMBEC coefficient accessed like pmbec_dict["V"]["R"]
+pmbec_dict = read_pmbec_coefficients(key_type="row")
+pmbec_matrix = dict_to_amino_acid_matrix(pmbec_dict)