Mapping UI is now working.

examples/Model use

@@ -8,31 +8,31 @@
       "targetHandle": "bundle"
     },
     {
-      "id": "Train/test split 1 Train model 3",
+      "id": "Train/test split 1 Define model 1",
       "source": "Train/test split 1",
       "sourceHandle": "output",
-      "target": "Train model 3",
+      "target": "Define model 1",
       "targetHandle": "bundle"
     },
     {
-      "id": "Model inference 1 View tables 1",
-      "source": "Model inference 1",
+      "id": "Define model 1 Train model 2",
+      "source": "Define model 1",
       "sourceHandle": "output",
-      "target": "View tables 1",
+      "target": "Train model 2",
       "targetHandle": "bundle"
     },
     {
-      "id": "Train/test split 1 Train model 1",
-      "source": "Train/test split 1",
+      "id": "Train model 2 Model inference 1",
+      "source": "Train model 2",
       "sourceHandle": "output",
-      "target": "Train model 1",
+      "target": "Model inference 1",
       "targetHandle": "bundle"
     },
     {
-      "id": "Train model 1 Model inference 1",
-      "source": "Train model 1",
+      "id": "Model inference 1 View tables 1",
+      "source": "Model inference 1",
       "sourceHandle": "output",
-      "target": "Model inference 1",
+      "target": "View tables 1",
       "targetHandle": "bundle"
     }
   ],
@@ -167,199 +167,6 @@
       "type": "basic",
       "width": 371.0
     },
-    {
-      "data": {
-        "__execution_delay": 0.0,
-        "collapsed": null,
-        "display": {
-          "dataframes": {
-            "df": {
-              "columns": [
-                "x",
-                "y"
-              ]
-            },
-            "df_test": {
-              "columns": [
-                "x",
-                "y"
-              ]
-            },
-            "df_train": {
-              "columns": [
-                "x",
-                "y"
-              ]
-            }
-          },
-          "other": {
-            "model": {
-              "model": {
-                "inputs": [
-                  "Input__embedding_1_x"
-                ],
-                "loss_inputs": [
-                  "Activation_2_x",
-                  "Input__label_1_y"
-                ],
-                "outputs": [
-                  "Activation_2_x"
-                ]
-              },
-              "type": "model"
-            }
-          },
-          "relations": []
-        },
-        "error": null,
-        "meta": {
-          "inputs": {
-            "bundle": {
-              "name": "bundle",
-              "position": "left",
-              "type": {
-                "type": "<class 'lynxkite_graph_analytics.core.Bundle'>"
-              }
-            }
-          },
-          "name": "Train model",
-          "outputs": {
-            "output": {
-              "name": "output",
-              "position": "right",
-              "type": {
-                "type": "None"
-              }
-            }
-          },
-          "params": {
-            "epochs": {
-              "default": 1.0,
-              "name": "epochs",
-              "type": {
-                "type": "<class 'int'>"
-              }
-            },
-            "input_mapping": {
-              "default": null,
-              "name": "input_mapping",
-              "type": {
-                "type": "<class 'str'>"
-              }
-            },
-            "model_workspace": {
-              "default": null,
-              "name": "model_workspace",
-              "type": {
-                "type": "<class 'str'>"
-              }
-            },
-            "save_as": {
-              "default": "model",
-              "name": "save_as",
-              "type": {
-                "type": "<class 'str'>"
-              }
-            }
-          },
-          "position": {
-            "x": 675.0,
-            "y": 144.0
-          },
-          "type": "basic"
-        },
-        "params": {
-          "epochs": "1000",
-          "input_mapping": "{\"map\": {\"Input__embedding_1_x\": {\"df\": \"df_train\", \"column\": \"x\"}, \"Input__label_1_y\": {\"df\": \"df_train\", \"column\": \"y\"  }}}",
-          "model_workspace": "Model definition",
-          "save_as": "model"
-        },
-        "status": "done",
-        "title": "Train model"
-      },
-      "dragHandle": ".bg-primary",
-      "height": 519.0,
-      "id": "Train model 3",
-      "position": {
-        "x": 722.5912720951791,
-        "y": -784.0614755260641
-      },
-      "type": "basic",
-      "width": 640.0
-    },
-    {
-      "data": {
-        "__execution_delay": 0.0,
-        "collapsed": null,
-        "display": null,
-        "error": "'Input__embedding_1_x'",
-        "meta": {
-          "inputs": {
-            "bundle": {
-              "name": "bundle",
-              "position": "left",
-              "type": {
-                "type": "<class 'lynxkite_graph_analytics.core.Bundle'>"
-              }
-            }
-          },
-          "name": "Model inference",
-          "outputs": {
-            "output": {
-              "name": "output",
-              "position": "right",
-              "type": {
-                "type": "None"
-              }
-            }
-          },
-          "params": {
-            "input_mapping": {
-              "default": "",
-              "name": "input_mapping",
-              "type": {
-                "type": "<class 'str'>"
-              }
-            },
-            "model_name": {
-              "default": "model",
-              "name": "model_name",
-              "type": {
-                "type": "<class 'str'>"
-              }
-            },
-            "output_mapping": {
-              "default": "",
-              "name": "output_mapping",
-              "type": {
-                "type": "<class 'str'>"
-              }
-            }
-          },
-          "position": {
-            "x": 506.0,
-            "y": 115.0
-          },
-          "type": "basic"
-        },
-        "params": {
-          "input_mapping": "{\"map\": {\"Input__embedding_1_x\": {\"df\": \"df_test\", \"column\": \"x\"}}}",
-          "model_name": "model",
-          "output_mapping": "{\"map\": {\"Activation_2_x\": {\"df\": \"df_test\", \"column\": \"predicted\"}}}"
-        },
-        "status": "done",
-        "title": "Model inference"
-      },
-      "dragHandle": ".bg-primary",
-      "height": 429.0,
-      "id": "Model inference 1",
-      "position": {
-        "x": 1445.5664910683593,
-        "y": 12.075943590382515
-      },
-      "type": "basic",
-      "width": 410.0
-    },
     {
       "data": {
         "display": {
@@ -780,54 +587,54 @@
               ],
               "data": [
                 [
-                  "[0.52046251 0.45887971 0.72169858 0.29517919]",
-                  "[1.52046251 1.45887971 1.72169852 1.29517913]",
-                  "[1.5168578624725342, 1.450861930847168, 1.7133464813232422, 1.3041404485702515]"
+                  "[0.4827103  0.10563457 0.98858833 0.82286644]",
+                  "[1.48271036 1.10563457 1.98858833 1.82286644]",
+                  "[1.4503304958343506, 1.1575173139572144, 1.996051549911499, 1.8715994358062744]"
                 ],
                 [
-                  "[0.78956431 0.87284744 0.06880784 0.03455889]",
-                  "[1.78956437 1.87284744 1.06880784 1.03455889]",
-                  "[1.7899272441864014, 1.829580307006836, 1.0702992677688599, 1.0265709161758423]"
+                  "[0.54914117 0.03810108 0.87531954 0.73044223]",
+                  "[1.54914117 1.03810108 1.87531948 1.73044229]",
+                  "[1.5054945945739746, 1.0799291133880615, 1.8784711360931396, 1.7745399475097656]"
                 ],
                 [
-                  "[0.49607176 0.1922397  0.46640229 0.78321403]",
-                  "[1.49607182 1.19223976 1.46640229 1.78321409]",
-                  "[1.4901000261306763, 1.193819284439087, 1.4632138013839722, 1.7822779417037964]"
+                  "[0.26661873 0.45946234 0.13510543 0.81294441]",
+                  "[1.26661873 1.4594624  1.13510537 1.81294441]",
+                  "[1.2893636226654053, 1.4012629985809326, 1.1273266077041626, 1.7483277320861816]"
                 ],
                 [
-                  "[0.49691743 0.61873293 0.90698647 0.94486356]",
-                  "[1.49691749 1.61873293 1.90698647 1.94486356]",
-                  "[1.4999868869781494, 1.6656270027160645, 1.9074199199676514, 1.9556759595870972]"
+                  "[0.9630242  0.76359051 0.24853623 0.76881069]",
+                  "[1.96302414 1.76359057 1.24853623 1.76881075]",
+                  "[1.9627504348754883, 1.808237075805664, 1.2652870416641235, 1.7989287376403809]"
                 ],
                 [
-                  "[0.59812403 0.78395379 0.0291847  0.81814629]",
-                  "[1.59812403 1.78395379 1.0291847  1.81814623]",
-                  "[1.6044235229492188, 1.7707669734954834, 1.0426081418991089, 1.7988944053649902]"
+                  "[0.52046251 0.45887971 0.72169858 0.29517919]",
+                  "[1.52046251 1.45887971 1.72169852 1.29517913]",
+                  "[1.4980013370513916, 1.4540209770202637, 1.7123804092407227, 1.3061636686325073]"
                 ],
                 [
-                  "[0.67447788 0.6125319  0.98007888 0.65968603]",
-                  "[1.67447782 1.6125319  1.98007894 1.65968609]",
-                  "[1.6721093654632568, 1.6624714136123657, 1.9726766347885132, 1.6813924312591553]"
+                  "[0.62569475 0.9881897  0.83639616 0.9828859 ]",
+                  "[1.62569475 1.9881897  1.83639622 1.98288584]",
+                  "[1.6632444858551025, 2.0650277137756348, 1.8643897771835327, 2.032055377960205]"
                 ],
                 [
-                  "[0.18720162 0.74115586 0.98626411 0.30355608]",
-                  "[1.18720162 1.74115586 1.98626411 1.30355608]",
-                  "[1.1961991786956787, 1.723442792892456, 1.9852817058563232, 1.3066248893737793]"
+                  "[0.95928186 0.84273899 0.71514636 0.38619852]",
+                  "[1.95928192 1.84273899 1.7151463  1.38619852]",
+                  "[1.943433403968811, 1.905775785446167, 1.727120280265808, 1.4508745670318604]"
                 ],
                 [
-                  "[0.74064726 0.4155122  0.09800029 0.49930882]",
-                  "[1.74064732 1.4155122  1.09800029 1.49930882]",
-                  "[1.7340764999389648, 1.3968157768249512, 1.0968588590621948, 1.493086814880371]"
+                  "[0.94516498 0.08422136 0.5608117  0.07652664]",
+                  "[1.94516492 1.08422136 1.56081176 1.07652664]",
+                  "[1.8603894710540771, 1.0990982055664062, 1.5473597049713135, 1.1221177577972412]"
                 ],
                 [
-                  "[0.70167565 0.26930219 0.5660674  0.61194974]",
-                  "[1.70167565 1.26930213 1.56606746 1.61194968]",
-                  "[1.691997766494751, 1.2865687608718872, 1.5571787357330322, 1.622729778289795]"
+                  "[0.32225502 0.16999388 0.05823922 0.9628762 ]",
+                  "[1.32225502 1.16999388 1.05823922 1.9628762 ]",
+                  "[1.3253933191299438, 1.1194703578948975, 1.0517100095748901, 1.903820514678955]"
                 ],
                 [
-                  "[0.90817457 0.89270043 0.38583666 0.66566533]",
-                  "[1.90817451 1.89270043 1.3858366  1.66566539]",
-                  "[1.9086859226226807, 1.924757719039917, 1.3887461423873901, 1.6714670658111572]"
+                  "[0.91730917 0.22574073 0.09591609 0.33056474]",
+                  "[1.91730917 1.22574067 1.09591603 1.33056474]",
+                  "[1.8627445697784424, 1.2110977172851562, 1.085142731666565, 1.3327856063842773]"
                 ]
               ]
             },
@@ -869,10 +676,6 @@
                   "[4.27091718e-01 4.89909172e-01 6.92297399e-01 2.57611275e-04]",
                   "[1.42709172 1.48990917 1.69229746 1.00025761]"
                 ],
-                [
-                  "[0.32225502 0.16999388 0.05823922 0.9628762 ]",
-                  "[1.32225502 1.16999388 1.05823922 1.9628762 ]"
-                ],
                 [
                   "[0.50783676 0.04156506 0.21984279 0.8454656 ]",
                   "[1.50783682 1.04156506 1.21984279 1.84546566]"
@@ -913,10 +716,6 @@
                   "[0.24388778 0.07268471 0.68350857 0.73431659]",
                   "[1.24388778 1.07268476 1.68350863 1.73431659]"
                 ],
-                [
-                  "[0.62569475 0.9881897  0.83639616 0.9828859 ]",
-                  "[1.62569475 1.9881897  1.83639622 1.98288584]"
-                ],
                 [
                   "[0.56922203 0.98222166 0.76851749 0.28615737]",
                   "[1.56922197 1.9822216  1.76851749 1.28615737]"
@@ -925,6 +724,10 @@
                   "[0.88776821 0.51636773 0.30333066 0.32230979]",
                   "[1.88776827 1.51636767 1.30333066 1.32230973]"
                 ],
+                [
+                  "[0.90817457 0.89270043 0.38583666 0.66566533]",
+                  "[1.90817451 1.89270043 1.3858366  1.66566539]"
+                ],
                 [
                   "[0.48507756 0.80808765 0.77162558 0.47834778]",
                   "[1.48507762 1.80808759 1.77162552 1.47834778]"
@@ -949,14 +752,6 @@
                   "[0.23942459 0.90487361 0.69337189 0.65089428]",
                   "[1.23942459 1.90487361 1.69337189 1.65089428]"
                 ],
-                [
-                  "[0.94516498 0.08422136 0.5608117  0.07652664]",
-                  "[1.94516492 1.08422136 1.56081176 1.07652664]"
-                ],
-                [
-                  "[0.26661873 0.45946234 0.13510543 0.81294441]",
-                  "[1.26661873 1.4594624  1.13510537 1.81294441]"
-                ],
                 [
                   "[0.30754459 0.77694583 0.09278506 0.38326019]",
                   "[1.30754459 1.77694583 1.09278512 1.38326025]"
@@ -965,14 +760,18 @@
                   "[0.27845025 0.32472342 0.82203609 0.77107543]",
                   "[1.27845025 1.32472348 1.82203603 1.77107549]"
                 ],
-                [
-                  "[0.4827103  0.10563457 0.98858833 0.82286644]",
-                  "[1.48271036 1.10563457 1.98858833 1.82286644]"
-                ],
                 [
                   "[0.98033333 0.97656083 0.38939917 0.81491041]",
                   "[1.98033333 1.97656083 1.38939917 1.81491041]"
                 ],
+                [
+                  "[0.74064726 0.4155122  0.09800029 0.49930882]",
+                  "[1.74064732 1.4155122  1.09800029 1.49930882]"
+                ],
+                [
+                  "[0.78956431 0.87284744 0.06880784 0.03455889]",
+                  "[1.78956437 1.87284744 1.06880784 1.03455889]"
+                ],
                 [
                   "[0.94221359 0.57740951 0.98649532 0.40934443]",
                   "[1.94221354 1.57740951 1.98649526 1.40934443]"
@@ -1021,6 +820,10 @@
                   "[0.08107251 0.2602725  0.18861133 0.44833237]",
                   "[1.08107257 1.2602725  1.18861127 1.44833231]"
                 ],
+                [
+                  "[0.59812403 0.78395379 0.0291847  0.81814629]",
+                  "[1.59812403 1.78395379 1.0291847  1.81814623]"
+                ],
                 [
                   "[0.93488538 0.73882395 0.37345302 0.0274905 ]",
                   "[1.93488538 1.73882389 1.37345302 1.0274905 ]"
@@ -1030,12 +833,12 @@
                   "[1.30631399 1.48311198 1.87847519 1.67559886]"
                 ],
                 [
-                  "[0.85566247 0.83362883 0.48424995 0.25265992]",
-                  "[1.85566247 1.83362889 1.48424995 1.25265992]"
+                  "[0.18720162 0.74115586 0.98626411 0.30355608]",
+                  "[1.18720162 1.74115586 1.98626411 1.30355608]"
                 ],
                 [
-                  "[0.95928186 0.84273899 0.71514636 0.38619852]",
-                  "[1.95928192 1.84273899 1.7151463  1.38619852]"
+                  "[0.85566247 0.83362883 0.48424995 0.25265992]",
+                  "[1.85566247 1.83362889 1.48424995 1.25265992]"
                 ],
                 [
                   "[0.32565445 0.90939188 0.07488042 0.13730896]",
@@ -1049,10 +852,6 @@
                   "[0.79905868 0.89367443 0.75429088 0.3190186 ]",
                   "[1.79905868 1.89367437 1.75429082 1.3190186 ]"
                 ],
-                [
-                  "[0.54914117 0.03810108 0.87531954 0.73044223]",
-                  "[1.54914117 1.03810108 1.87531948 1.73044229]"
-                ],
                 [
                   "[0.67418337 0.79634351 0.23229051 0.71345252]",
                   "[1.67418337 1.79634356 1.23229051 1.71345258]"
@@ -1090,12 +889,12 @@
                   "[1.60809326 1.5656302  1.32107437 1.72599435]"
                 ],
                 [
-                  "[0.47963417 0.81818312 0.48720706 0.49339259]",
-                  "[1.47963417 1.81818318 1.48720706 1.49339259]"
+                  "[0.67447788 0.6125319  0.98007888 0.65968603]",
+                  "[1.67447782 1.6125319  1.98007894 1.65968609]"
                 ],
                 [
-                  "[0.9630242  0.76359051 0.24853623 0.76881069]",
-                  "[1.96302414 1.76359057 1.24853623 1.76881075]"
+                  "[0.47963417 0.81818312 0.48720706 0.49339259]",
+                  "[1.47963417 1.81818318 1.48720706 1.49339259]"
                 ],
                 [
                   "[0.60609657 0.96257663 0.19292736 0.95702219]",
@@ -1105,6 +904,10 @@
                   "[0.80654246 0.08253473 0.74478531 0.71257162]",
                   "[1.8065424  1.08253479 1.74478531 1.71257162]"
                 ],
+                [
+                  "[0.70167565 0.26930219 0.5660674  0.61194974]",
+                  "[1.70167565 1.26930213 1.56606746 1.61194968]"
+                ],
                 [
                   "[0.76933283 0.86241865 0.44114518 0.65644735]",
                   "[1.76933289 1.86241865 1.44114518 1.65644741]"
@@ -1122,8 +925,8 @@
                   "[1.12024069 1.21342516 1.56858408 1.58644271]"
                 ],
                 [
-                  "[0.91730917 0.22574073 0.09591609 0.33056474]",
-                  "[1.91730917 1.22574067 1.09591603 1.33056474]"
+                  "[0.49691743 0.61873293 0.90698647 0.94486356]",
+                  "[1.49691749 1.61873293 1.90698647 1.94486356]"
                 ],
                 [
                   "[0.6032477  0.83361369 0.18538666 0.19108021]",
@@ -1137,6 +940,10 @@
                   "[0.37959969 0.42820001 0.10690689 0.96353984]",
                   "[1.37959969 1.42820001 1.10690689 1.96353984]"
                 ],
+                [
+                  "[0.49607176 0.1922397  0.46640229 0.78321403]",
+                  "[1.49607182 1.19223976 1.46640229 1.78321409]"
+                ],
                 [
                   "[0.40234613 0.54987347 0.49542785 0.54153186]",
                   "[1.40234613 1.54987347 1.49542785 1.5415318 ]"
@@ -1201,7 +1008,7 @@
             }
           },
           "other": {
-            "model": "ModelConfig(model=Sequential(\n  (0) - Linear(in_features=4, out_features=4, bias=True): Input__embedding_1_x -> Linear_1_x\n  (1) - <function leaky_relu at 0x710fc8f0fba0>: Linear_1_x -> Activation_2_x\n  (2) - Identity(): Activation_2_x -> Activation_2_x\n), model_inputs=['Input__embedding_1_x'], model_outputs=['Activation_2_x'], loss_inputs=['Activation_2_x', 'Input__label_1_y'], loss=Sequential(\n  (0) - <function mse_loss at 0x710fc8f316c0>: Activation_2_x, Input__label_1_y -> MSE_loss_1_loss\n  (1) - Identity(): MSE_loss_1_loss -> loss\n), optimizer=SGD (\nParameter Group 0\n    dampening: 0\n    differentiable: False\n    foreach: None\n    fused: None\n    lr: 0.1\n    maximize: False\n    momentum: 0\n    nesterov: False\n    weight_decay: 0\n))"
+            "model": "ModelConfig(model=Sequential(\n  (0) - Linear(in_features=4, out_features=4, bias=True): Input__embedding_1_x -> Linear_1_x\n  (1) - <function leaky_relu at 0x731d1381bba0>: Linear_1_x -> Activation_2_x\n  (2) - Identity(): Activation_2_x -> Activation_2_x\n), model_inputs=['Input__embedding_1_x'], model_outputs=['Activation_2_x'], loss_inputs=['Input__label_1_y', 'Activation_2_x'], loss=Sequential(\n  (0) - <function mse_loss at 0x731d138216c0>: Activation_2_x, Input__label_1_y -> MSE_loss_1_loss\n  (1) - Identity(): MSE_loss_1_loss -> loss\n), optimizer=SGD (\nParameter Group 0\n    dampening: 0\n    differentiable: False\n    foreach: None\n    fused: None\n    lr: 0.1\n    maximize: False\n    momentum: 0\n    nesterov: False\n    weight_decay: 0\n), source_workspace=None, trained=True)"
           },
           "relations": []
         },
@@ -1227,27 +1034,23 @@
               }
             }
           },
-          "position": {
-            "x": 471.0,
-            "y": 424.0
-          },
           "type": "table_view"
         },
         "params": {
           "limit": 100.0
         },
-        "status": "planned",
+        "status": "done",
         "title": "View tables"
       },
       "dragHandle": ".bg-primary",
-      "height": 600.0,
+      "height": 711.0,
       "id": "View tables 1",
       "position": {
-        "x": 2017.4630208327735,
-        "y": -223.54449620081252
+        "x": 2900.3734843758352,
+        "y": -128.41609608842867
       },
       "type": "table_view",
-      "width": 582.0
+      "width": 836.0
     },
     {
       "data": {
@@ -1277,22 +1080,125 @@
           "other": {
             "model": {
               "model": {
-                "inputs": [],
+                "inputs": [
+                  "Input__embedding_1_x"
+                ],
                 "loss_inputs": [
-                  "Activation_2_x",
-                  "Input__label_1_y"
+                  "Input__label_1_y",
+                  "Activation_2_x"
                 ],
                 "outputs": [
-                  "Activation_2_x",
-                  "Input__label_1_y"
-                ]
+                  "Activation_2_x"
+                ],
+                "trained": false
               },
               "type": "model"
             }
           },
           "relations": []
         },
-        "error": "Mapping is unset.",
+        "error": null,
+        "meta": {
+          "inputs": {
+            "bundle": {
+              "name": "bundle",
+              "position": "left",
+              "type": {
+                "type": "<class 'lynxkite_graph_analytics.core.Bundle'>"
+              }
+            }
+          },
+          "name": "Define model",
+          "outputs": {
+            "output": {
+              "name": "output",
+              "position": "right",
+              "type": {
+                "type": "None"
+              }
+            }
+          },
+          "params": {
+            "model_workspace": {
+              "default": null,
+              "name": "model_workspace",
+              "type": {
+                "type": "<class 'str'>"
+              }
+            },
+            "save_as": {
+              "default": "model",
+              "name": "save_as",
+              "type": {
+                "type": "<class 'str'>"
+              }
+            }
+          },
+          "type": "basic"
+        },
+        "params": {
+          "model_workspace": "Model definition",
+          "save_as": "model"
+        },
+        "status": "done",
+        "title": "Define model"
+      },
+      "dragHandle": ".bg-primary",
+      "height": 537.0,
+      "id": "Define model 1",
+      "position": {
+        "x": 795.0,
+        "y": -45.0
+      },
+      "type": "basic",
+      "width": 498.0
+    },
+    {
+      "data": {
+        "__execution_delay": 0.0,
+        "collapsed": null,
+        "display": {
+          "dataframes": {
+            "df": {
+              "columns": [
+                "x",
+                "y"
+              ]
+            },
+            "df_test": {
+              "columns": [
+                "x",
+                "y"
+              ]
+            },
+            "df_train": {
+              "columns": [
+                "x",
+                "y"
+              ]
+            }
+          },
+          "other": {
+            "model": {
+              "model": {
+                "inputs": [
+                  "Input__embedding_1_x"
+                ],
+                "loss_inputs": [
+                  "Input__label_1_y",
+                  "Activation_2_x"
+                ],
+                "outputs": [
+                  "Activation_2_x"
+                ],
+                "trained": true
+              },
+              "type": "model"
+            }
+          },
+          "relations": []
+        },
+        "error": null,
         "meta": {
           "inputs": {
             "bundle": {
@@ -1325,48 +1231,154 @@
               "default": null,
               "name": "input_mapping",
               "type": {
-                "type": "<class 'lynxkite_graph_analytics.pytorch_model_ops.ModelMapping'>"
+                "type": "<class 'lynxkite_graph_analytics.lynxkite_ops.ModelTrainingInputMapping'>"
               }
             },
-            "model_workspace": {
-              "default": null,
-              "name": "model_workspace",
+            "model_name": {
+              "default": "model",
+              "name": "model_name",
               "type": {
                 "type": "<class 'str'>"
               }
+            }
+          },
+          "position": {
+            "x": 666.0,
+            "y": 492.0
+          },
+          "type": "basic"
+        },
+        "params": {
+          "epochs": "1000",
+          "input_mapping": "{\"map\":{\"Input__embedding_1_x\":{\"column\":\"x\",\"df\":\"df_train\"},\"Input__label_1_y\":{\"column\":\"y\",\"df\":\"df_train\"}}}",
+          "model_name": "model"
+        },
+        "status": "done",
+        "title": "Train model"
+      },
+      "dragHandle": ".bg-primary",
+      "height": 604.0,
+      "id": "Train model 2",
+      "position": {
+        "x": 1399.5245787239226,
+        "y": -19.196202428593544
+      },
+      "type": "basic",
+      "width": 586.0
+    },
+    {
+      "data": {
+        "__execution_delay": 0.0,
+        "collapsed": null,
+        "display": {
+          "dataframes": {
+            "df": {
+              "columns": [
+                "x",
+                "y"
+              ]
             },
-            "save_as": {
+            "df_test": {
+              "columns": [
+                "predicted",
+                "x",
+                "y"
+              ]
+            },
+            "df_train": {
+              "columns": [
+                "x",
+                "y"
+              ]
+            }
+          },
+          "other": {
+            "model": {
+              "model": {
+                "inputs": [
+                  "Input__embedding_1_x"
+                ],
+                "loss_inputs": [
+                  "Input__label_1_y",
+                  "Activation_2_x"
+                ],
+                "outputs": [
+                  "Activation_2_x"
+                ],
+                "trained": true
+              },
+              "type": "model"
+            }
+          },
+          "relations": []
+        },
+        "error": null,
+        "meta": {
+          "inputs": {
+            "bundle": {
+              "name": "bundle",
+              "position": "left",
+              "type": {
+                "type": "<class 'lynxkite_graph_analytics.core.Bundle'>"
+              }
+            }
+          },
+          "name": "Model inference",
+          "outputs": {
+            "output": {
+              "name": "output",
+              "position": "right",
+              "type": {
+                "type": "None"
+              }
+            }
+          },
+          "params": {
+            "input_mapping": {
+              "default": null,
+              "name": "input_mapping",
+              "type": {
+                "type": "<class 'lynxkite_graph_analytics.lynxkite_ops.ModelInferenceInputMapping'>"
+              }
+            },
+            "model_name": {
               "default": "model",
-              "name": "save_as",
+              "name": "model_name",
               "type": {
                 "type": "<class 'str'>"
               }
+            },
+            "output_mapping": {
+              "default": null,
+              "name": "output_mapping",
+              "type": {
+                "type": "<class 'lynxkite_graph_analytics.lynxkite_ops.ModelOutputMapping'>"
+              }
             }
           },
           "position": {
-            "x": 723.0,
-            "y": 370.0
+            "x": 934.0,
+            "y": 167.0
           },
           "type": "basic"
         },
         "params": {
-          "epochs": "2",
-          "input_mapping": "{\"map\":{\"Activation_2_x\":{\"df\":\"df_train\"},\"Input__label_1_y\":{\"df\":\"df_train\",\"column\":\"y\"}}}",
-          "model_workspace": "Model definition",
-          "save_as": "model"
+          "input_mapping": "{\"map\":{\"Input__embedding_1_x\":{\"column\":\"x\",\"df\":\"df_test\"}}}",
+          "model_name": "model",
+          "output_mapping": "{\"map\":{\"Activation_2_x\":{\"column\":\"predicted\",\"df\":\"df_test\"}}}"
         },
         "status": "done",
-        "title": "Train model"
+        "title": "Model inference"
       },
       "dragHandle": ".bg-primary",
-      "height": 473.0,
-      "id": "Train model 1",
+      "height": 893.0,
+      "id": "Model inference 1",
       "position": {
-        "x": 712.1212754578014,
-        "y": 42.33722689912529
+        "x": 2181.718373860645,
+        "y": -69.44701793295484
       },
       "type": "basic",
-      "width": 577.0
+      "width": 529.0
     }
   ]
 }

lynxkite-app/web/src/workspace/nodes/NodeParameter.tsx

@@ -2,17 +2,54 @@
 import ArrowsHorizontal from "~icons/tabler/arrows-horizontal.jsx";
 
 const BOOLEAN = "<class 'bool'>";
-const MODEL_MAPPING =
-  "<class 'lynxkite_graph_analytics.pytorch_model_ops.ModelMapping'>";
+const MODEL_TRAINING_INPUT_MAPPING =
+  "<class 'lynxkite_graph_analytics.lynxkite_ops.ModelTrainingInputMapping'>";
+const MODEL_INFERENCE_INPUT_MAPPING =
+  "<class 'lynxkite_graph_analytics.lynxkite_ops.ModelInferenceInputMapping'>";
+const MODEL_OUTPUT_MAPPING =
+  "<class 'lynxkite_graph_analytics.lynxkite_ops.ModelOutputMapping'>";
 function ParamName({ name }: { name: string }) {
   return (
     <span className="param-name bg-base-200">{name.replace(/_/g, " ")}</span>
   );
 }
 
-function getModelBindings(data: any): string[] {
+function Input({
+  value,
+  onChange,
+}: {
+  value: string;
+  onChange: (value: string, options?: { delay: number }) => void;
+}) {
+  return (
+    <input
+      className="input input-bordered w-full"
+      value={value || ""}
+      onChange={(evt) => onChange(evt.currentTarget.value, { delay: 2 })}
+      onBlur={(evt) => onChange(evt.currentTarget.value, { delay: 0 })}
+      onKeyDown={(evt) =>
+        evt.code === "Enter" && onChange(evt.currentTarget.value, { delay: 0 })
+      }
+    />
+  );
+}
+
+function getModelBindings(
+  data: any,
+  variant: "training input" | "inference input" | "output",
+): string[] {
   function bindingsOfModel(m: any): string[] {
-    return [...m.inputs, ...m.outputs, ...m.loss_inputs];
+    switch (variant) {
+      case "training input":
+        return [
+          ...m.inputs,
+          ...m.loss_inputs.filter((i: string) => !m.outputs.includes(i)),
+        ];
+      case "inference input":
+        return m.inputs;
+      case "output":
+        return m.outputs;
+    }
   }
   const bindings = new Set<string>();
   const other = data?.display?.other ?? data?.display?.value?.other ?? {};
@@ -31,19 +68,19 @@ function getModelBindings(data: any): string[] {
 function parseJsonOrEmpty(json: string): object {
   try {
     const j = JSON.parse(json);
-    if (typeof j === "object") {
+    if (j !== null && typeof j === "object") {
       return j;
     }
   } catch (e) {}
   return {};
 }
 
-function ModelMapping({ value, onChange, data }: any) {
+function ModelMapping({ value, onChange, data, variant }: any) {
   const v: any = parseJsonOrEmpty(value);
   v.map ??= {};
   const dfs =
     data?.display?.dataframes ?? data?.display?.value?.dataframes ?? {};
-  const bindings = getModelBindings(data);
+  const bindings = getModelBindings(data, variant);
   return (
     <table className="model-mapping-param">
       <tbody>
@@ -63,12 +100,12 @@ function ModelMapping({ value, onChange, data }: any) {
                   value={v.map?.[binding]?.df}
                   onChange={(evt) => {
                     const df = evt.currentTarget.value;
-                    if (df === "unbound") {
+                    if (df === "") {
                       const map = { ...v.map, [binding]: undefined };
                       onChange(JSON.stringify({ map }));
                     } else {
                       const columnSpec = {
-                        column: dfs[df][0],
+                        column: dfs[df].columns[0],
                         ...(v.map?.[binding] || {}),
                         df,
                       };
@@ -77,9 +114,7 @@ function ModelMapping({ value, onChange, data }: any) {
                     }
                   }}
                 >
-                  <option key="unbound" value="unbound">
-                    unbound
-                  </option>
+                  <option key="" value="" />
                   {Object.keys(dfs).map((df: string) => (
                     <option key={df} value={df}>
                       {df}
@@ -88,22 +123,39 @@ function ModelMapping({ value, onChange, data }: any) {
                 </select>
               </td>
               <td>
-                <select
-                  className="select select-ghost"
-                  value={v.map?.[binding]?.column}
-                  onChange={(evt) => {
-                    const column = evt.currentTarget.value;
-                    const columnSpec = { ...(v.map?.[binding] || {}), column };
-                    const map = { ...v.map, [binding]: columnSpec };
-                    onChange(JSON.stringify({ map }));
-                  }}
-                >
-                  {dfs[v.map?.[binding]?.df]?.columns.map((col: string) => (
-                    <option key={col} value={col}>
-                      {col}
-                    </option>
-                  ))}
-                </select>
+                {variant === "output" ? (
+                  <Input
+                    value={v.map?.[binding]?.column}
+                    onChange={(column, options) => {
+                      const columnSpec = {
+                        ...(v.map?.[binding] || {}),
+                        column,
+                      };
+                      const map = { ...v.map, [binding]: columnSpec };
+                      onChange(JSON.stringify({ map }), options);
+                    }}
+                  />
+                ) : (
+                  <select
+                    className="select select-ghost"
+                    value={v.map?.[binding]?.column}
+                    onChange={(evt) => {
+                      const column = evt.currentTarget.value;
+                      const columnSpec = {
+                        ...(v.map?.[binding] || {}),
+                        column,
+                      };
+                      const map = { ...v.map, [binding]: columnSpec };
+                      onChange(JSON.stringify({ map }));
+                    }}
+                  >
+                    {dfs[v.map?.[binding]?.df]?.columns.map((col: string) => (
+                      <option key={col} value={col}>
+                        {col}
+                      </option>
+                    ))}
+                  </select>
+                )}
               </td>
             </tr>
           ))
@@ -178,25 +230,41 @@ export default function NodeParameter({
             {name.replace(/_/g, " ")}
           </label>
         </div>
-      ) : meta?.type?.type === MODEL_MAPPING ? (
+      ) : meta?.type?.type === MODEL_TRAINING_INPUT_MAPPING ? (
         <>
           <ParamName name={name} />
-          <ModelMapping value={value} data={data} onChange={onChange} />
+          <ModelMapping
+            value={value}
+            data={data}
+            variant="training input"
+            onChange={onChange}
+          />
         </>
-      ) : (
+      ) : meta?.type?.type === MODEL_INFERENCE_INPUT_MAPPING ? (
         <>
           <ParamName name={name} />
-          <input
-            className="input input-bordered w-full"
-            value={value || ""}
-            onChange={(evt) => onChange(evt.currentTarget.value, { delay: 2 })}
-            onBlur={(evt) => onChange(evt.currentTarget.value, { delay: 0 })}
-            onKeyDown={(evt) =>
-              evt.code === "Enter" &&
-              onChange(evt.currentTarget.value, { delay: 0 })
-            }
+          <ModelMapping
+            value={value}
+            data={data}
+            variant="inference input"
+            onChange={onChange}
+          />
+        </>
+      ) : meta?.type?.type === MODEL_OUTPUT_MAPPING ? (
+        <>
+          <ParamName name={name} />
+          <ModelMapping
+            value={value}
+            data={data}
+            variant="output"
+            onChange={onChange}
           />
         </>
+      ) : (
+        <>
+          <ParamName name={name} />
+          <Input value={value} onChange={onChange} />
+        </>
       )}
     </label>
   );

lynxkite-graph-analytics/src/lynxkite_graph_analytics/core.py

@@ -204,8 +204,8 @@ def _execute_node(node, ws, catalog, outputs):
         for edge in ws.edges
         if edge.target == node.id
     }
+    # Convert inputs types to match operation signature.
     try:
-        # Convert inputs types to match operation signature.
         inputs = []
         for p in op.inputs.values():
             if p.name not in input_map:
@@ -219,13 +219,24 @@ def _execute_node(node, ws, catalog, outputs):
             elif p.type == Bundle and isinstance(x, pd.DataFrame):
                 x = Bundle.from_df(x)
             inputs.append(x)
-        result = op(*inputs, **params)
     except Exception as e:
         if os.environ.get("LYNXKITE_LOG_OP_ERRORS"):
             traceback.print_exc()
         node.publish_error(e)
         return
-    outputs[node.id] = result.output
+    # Execute op.
+    try:
+        result = op(*inputs, **params)
+    except Exception as e:
+        if os.environ.get("LYNXKITE_LOG_OP_ERRORS"):
+            traceback.print_exc()
+        result = ops.Result(error=str(e))
+        # On error, just output the first input. This helps reduce the errors on the frontend,
+        # and it lets boxes easily access things from their inputs on the UI, even in error state.
+        if inputs:
+            result.output = inputs[0]
+    if result.output is not None:
+        outputs[node.id] = result.output
     node.publish_result(result)
 
 

lynxkite-graph-analytics/src/lynxkite_graph_analytics/lynxkite_ops.py

@@ -363,31 +363,60 @@ def biomedical_foundation_graph(*, filter_nodes: str):
     return None
 
 
-@op("Train model")
-def train_model(
+@op("Define model")
+def define_model(
     bundle: core.Bundle,
     *,
     model_workspace: str,
-    input_mapping: pytorch_model_ops.ModelMapping,
-    epochs: int = 1,
     save_as: str = "model",
 ):
     """Trains the selected model on the selected dataset. Most training parameters are set in the model definition."""
     assert model_workspace, "Model workspace is unset."
-    print(f"input_mapping: {input_mapping}")
     ws = load_ws(model_workspace)
-    inputs = (
-        pytorch_model_ops.to_tensors(bundle, input_mapping) if input_mapping else {}
-    )
-    m = pytorch_model_ops.build_model(ws, inputs)
+    # Build the model without inputs, to get its interface.
+    m = pytorch_model_ops.build_model(ws, {})
+    m.source_workspace = model_workspace
     bundle = bundle.copy()
     bundle.other[save_as] = m
-    if input_mapping is None:
-        return ops.Result(bundle, error="Mapping is unset.")
+    return bundle
+
+
+# These contain the same mapping, but they get different UIs.
+# For inputs, you select existing columns. For outputs, you can create new columns.
+class ModelInferenceInputMapping(pytorch_model_ops.ModelMapping):
+    pass
+
+
+class ModelTrainingInputMapping(pytorch_model_ops.ModelMapping):
+    pass
+
+
+class ModelOutputMapping(pytorch_model_ops.ModelMapping):
+    pass
+
+
+@op("Train model")
+def train_model(
+    bundle: core.Bundle,
+    *,
+    model_name: str = "model",
+    input_mapping: ModelTrainingInputMapping,
+    epochs: int = 1,
+):
+    """Trains the selected model on the selected dataset. Most training parameters are set in the model definition."""
+    m = bundle.other[model_name].copy()
+    inputs = pytorch_model_ops.to_tensors(bundle, input_mapping)
+    if not m.trained and m.source_workspace:
+        # Rebuild the model for the correct inputs.
+        ws = load_ws(m.source_workspace)
+        m = pytorch_model_ops.build_model(ws, inputs)
     t = tqdm(range(epochs), desc="Training model")
     for _ in t:
         loss = m.train(inputs)
         t.set_postfix({"loss": loss})
+    m.trained = True
+    bundle = bundle.copy()
+    bundle.other[model_name] = m
     return bundle
 
 
@@ -396,13 +425,14 @@ def model_inference(
     bundle: core.Bundle,
     *,
     model_name: str = "model",
-    input_mapping: pytorch_model_ops.ModelMapping,
-    output_mapping: pytorch_model_ops.ModelMapping,
+    input_mapping: ModelInferenceInputMapping,
+    output_mapping: ModelOutputMapping,
 ):
     """Executes a trained model."""
     if input_mapping is None or output_mapping is None:
         return ops.Result(bundle, error="Mapping is unset.")
     m = bundle.other[model_name]
+    assert m.trained, "The model is not trained."
     inputs = pytorch_model_ops.to_tensors(bundle, input_mapping)
     outputs = m.inference(inputs)
     bundle = bundle.copy()

lynxkite-graph-analytics/src/lynxkite_graph_analytics/pytorch_model_ops.py

@@ -1,13 +1,15 @@
 """Boxes for defining PyTorch models."""
 
+import copy
 import graphlib
+import types
 
 import pydantic
 from lynxkite.core import ops, workspace
 from lynxkite.core.ops import Parameter as P
 import torch
 import torch_geometric as pyg
-from dataclasses import dataclass
+import dataclasses
 from . import core
 
 ENV = "PyTorch model"
@@ -125,9 +127,9 @@ ops.register_passive_op(
 )
 
 
-def _to_id(s: str) -> str:
+def _to_id(*strings: str) -> str:
     """Replaces all non-alphanumeric characters with underscores."""
-    return "".join(c if c.isalnum() else "_" for c in s)
+    return "_".join("".join(c if c.isalnum() else "_" for c in s) for s in strings)
 
 
 class ColumnSpec(pydantic.BaseModel):
@@ -139,7 +141,7 @@ class ModelMapping(pydantic.BaseModel):
     map: dict[str, ColumnSpec]
 
 
-@dataclass
+@dataclasses.dataclass
 class ModelConfig:
     model: torch.nn.Module
     model_inputs: list[str]
@@ -147,6 +149,8 @@ class ModelConfig:
     loss_inputs: list[str]
     loss: torch.nn.Module
     optimizer: torch.optim.Optimizer
+    source_workspace: str | None = None
+    trained: bool = False
 
     def _forward(self, inputs: dict[str, torch.Tensor]) -> dict[str, torch.Tensor]:
         model_inputs = [inputs[i] for i in self.model_inputs]
@@ -176,8 +180,8 @@ class ModelConfig:
 
     def copy(self):
         """Returns a copy of the model."""
-        c = super().copy()
-        c.model = self.model.copy()
+        c = dataclasses.replace(self)
+        c.model = copy.deepcopy(self.model)
         return c
 
     def default_display(self):
@@ -187,6 +191,7 @@ class ModelConfig:
                 "inputs": self.model_inputs,
                 "outputs": self.model_outputs,
                 "loss_inputs": self.loss_inputs,
+                "trained": self.trained,
             },
         }
 
@@ -206,13 +211,17 @@ def build_model(
     assert len(optimizers) == 1, f"More than one optimizer found: {optimizers}"
     [optimizer] = optimizers
     dependencies = {n.id: [] for n in ws.nodes}
-    edges = {}
+    in_edges = {}
+    out_edges = {}
     # TODO: Dissolve repeat boxes here.
     for e in ws.edges:
         dependencies[e.target].append(e.source)
-        edges.setdefault((e.target, e.targetHandle), []).append(
+        in_edges.setdefault(e.target, {}).setdefault(e.targetHandle, []).append(
             (e.source, e.sourceHandle)
         )
+        out_edges.setdefault(e.source, {}).setdefault(e.sourceHandle, []).append(
+            (e.target, e.targetHandle)
+        )
     sizes = {}
     for k, i in inputs.items():
         sizes[k] = i.shape[-1]
@@ -221,8 +230,10 @@ def build_model(
     loss_layers = []
     in_loss = set()
     cfg = {}
-    loss_inputs = set()
-    used_inputs = set()
+    used_in_model = set()
+    made_in_model = set()
+    used_in_loss = set()
+    made_in_loss = set()
     for node_id in ts.static_order():
         node = nodes[node_id]
         t = node.data.title
@@ -231,51 +242,62 @@ def build_model(
         for b in dependencies[node_id]:
             if b in in_loss:
                 in_loss.add(node_id)
+        if "loss" in t:
+            in_loss.add(node_id)
+        inputs = {}
+        for n in in_edges.get(node_id, []):
+            for b, h in in_edges[node_id][n]:
+                i = _to_id(b, h)
+                inputs[n] = i
+                if node_id in in_loss:
+                    used_in_loss.add(i)
+                else:
+                    used_in_model.add(i)
+        outputs = {}
+        for out in out_edges.get(node_id, []):
+            i = _to_id(node_id, out)
+            outputs[out] = i
+            if inputs:  # Nodes with no inputs are input nodes. Their outputs are not "made" by us.
+                if node_id in in_loss:
+                    made_in_loss.add(i)
+                else:
+                    made_in_model.add(i)
+        inputs = types.SimpleNamespace(**inputs)
+        outputs = types.SimpleNamespace(**outputs)
         ls = loss_layers if node_id in in_loss else layers
-        nid = _to_id(node_id)
         match t:
             case "Linear":
-                [(ib, ih)] = edges[node_id, "x"]
-                i = _to_id(ib) + "_" + ih
-                used_inputs.add(i)
-                isize = sizes[i]
+                isize = sizes.get(inputs.x, 1)
                 osize = isize if p["output_dim"] == "same" else int(p["output_dim"])
-                ls.append((torch.nn.Linear(isize, osize), f"{i} -> {nid}_x"))
-                sizes[f"{nid}_x"] = osize
+                ls.append((torch.nn.Linear(isize, osize), f"{inputs.x} -> {outputs.x}"))
+                sizes[outputs.x] = osize
             case "Activation":
-                [(ib, ih)] = edges[node_id, "x"]
-                i = _to_id(ib) + "_" + ih
-                used_inputs.add(i)
                 f = getattr(
                     torch.nn.functional, p["type"].name.lower().replace(" ", "_")
                 )
-                ls.append((f, f"{i} -> {nid}_x"))
-                sizes[f"{nid}_x"] = sizes[i]
+                ls.append((f, f"{inputs.x} -> {outputs.x}"))
+                sizes[outputs.x] = sizes.get(inputs.x, 1)
             case "MSE loss":
-                [(xb, xh)] = edges[node_id, "x"]
-                xi = _to_id(xb) + "_" + xh
-                [(yb, yh)] = edges[node_id, "y"]
-                yi = _to_id(yb) + "_" + yh
-                loss_inputs.add(xi)
-                loss_inputs.add(yi)
-                in_loss.add(node_id)
-                loss_layers.append(
-                    (torch.nn.functional.mse_loss, f"{xi}, {yi} -> {nid}_loss")
+                ls.append(
+                    (
+                        torch.nn.functional.mse_loss,
+                        f"{inputs.x}, {inputs.y} -> {outputs.loss}",
+                    )
                 )
-    cfg["model_inputs"] = list(used_inputs & inputs.keys())
-    cfg["model_outputs"] = list(loss_inputs - inputs.keys())
-    cfg["loss_inputs"] = list(loss_inputs)
+    cfg["model_inputs"] = list(used_in_model - made_in_model)
+    cfg["model_outputs"] = list(made_in_model & used_in_loss)
+    cfg["loss_inputs"] = list(used_in_loss - made_in_loss)
     # Make sure the trained output is output from the last model layer.
     outputs = ", ".join(cfg["model_outputs"])
     layers.append((torch.nn.Identity(), f"{outputs} -> {outputs}"))
     # Create model.
-    cfg["model"] = pyg.nn.Sequential(", ".join(used_inputs & inputs.keys()), layers)
+    cfg["model"] = pyg.nn.Sequential(", ".join(cfg["model_inputs"]), layers)
     # Make sure the loss is output from the last loss layer.
-    [(lossb, lossh)] = edges[optimizer, "loss"]
-    lossi = _to_id(lossb) + "_" + lossh
+    [(lossb, lossh)] = in_edges[optimizer]["loss"]
+    lossi = _to_id(lossb, lossh)
     loss_layers.append((torch.nn.Identity(), f"{lossi} -> loss"))
     # Create loss function.
-    cfg["loss"] = pyg.nn.Sequential(", ".join(loss_inputs), loss_layers)
+    cfg["loss"] = pyg.nn.Sequential(", ".join(cfg["loss_inputs"]), loss_layers)
     assert not list(cfg["loss"].parameters()), (
         f"loss should have no parameters: {list(cfg['loss'].parameters())}"
     )
@@ -287,9 +309,14 @@ def build_model(
     return ModelConfig(**cfg)
 
 
-def to_tensors(b: core.Bundle, m: ModelMapping) -> dict[str, torch.Tensor]:
-    """Converts a tensor to the correct type for PyTorch."""
+def to_tensors(b: core.Bundle, m: ModelMapping | None) -> dict[str, torch.Tensor]:
+    """Converts a tensor to the correct type for PyTorch. Ignores missing mappings."""
+    if m is None:
+        return {}
     tensors = {}
     for k, v in m.map.items():
-        tensors[k] = torch.tensor(b.dfs[v.df][v.column].to_list(), dtype=torch.float32)
+        if v.df in b.dfs and v.column in b.dfs[v.df]:
+            tensors[k] = torch.tensor(
+                b.dfs[v.df][v.column].to_list(), dtype=torch.float32
+            )
     return tensors