Epic/sof 2008 - ML POC, PD [DO NOT MERGE]

example/job.json

@@ -1,9 +1,11 @@
 {
     "status": "pre-submission",
     "mode": "normal",
-    "_material": {
-        "...": "include(job/_material.json)"
-    },
+    "_materials": [
+        {
+            "...": "include(job/_material.json)"
+        }
+    ],
     "compute": {
         "...": "include(job/compute.json)"
     },
@@ -24,5 +26,8 @@
     },
     "_project": {
         "...": "include(job/_project.json)"
+    },
+    "api": {
+        "...": "include(job/_api.json)"
     }
 }

example/job/_api.json

@@ -0,0 +1,4 @@
+{
+    "userId": "jnLJpf9vJKYtFoQxc",
+    "authToken": "BvsTQ3xv-c4laO4Po3PQWsQty4GVwtDL_f2KXv6uFxz"
+}

example/job/input.json

@@ -0,0 +1,3 @@
+{
+    "...": "include(theory/ml/input.json)"
+}

example/job/software/ml/linear/workflow/fit_predict_band_gap.json

@@ -0,0 +1,119 @@
+{
+    "_id": "ML bandgap prediction",
+    "title": "SiGe alloy bandgap prediction",
+    "titleSlug": "Template for ML POC SiGe alloy bandgap prediction",
+    "compute": {
+        "ppn": 1,
+        "queue": "OR",
+        "timeLimit": "1:00:00",
+        "notify": "n",
+        "nodes": 1,
+        "arguments": {}
+    },
+    "model": {
+        "method": {
+            "workflow": {
+                "name": "sige alloy bandgap prediciton using least squares regression",
+                "app": "machine_learning",
+                "units": [
+                    {
+                        "flowchartId": "1",
+                        "head": true,
+                        "type": "execution",
+                        "name": "fit_MLmethod",
+                        "next": "2",
+                        "execution": {
+                            "app": {
+                                "name": "ml",
+                                "exec": "rupy",
+                                "version": "0.1.0"
+                            },
+                            "preProcessors": [],
+                            "input": [
+                                {
+                                    "fit": {
+                                        "type": "linear",
+                                        "subtype": "least_squares",
+                                        "arguments": {
+                                            "sampleWeight": true,
+                                            "trainDataSize": 0.75
+                                        }
+                                    },
+                                    "targetProperties": [
+                                        "band_gaps"
+                                    ],
+                                    "features": [
+                                        "elemental_composition",
+                                        "density"
+                                    ],
+                                    "exabyteIds": [
+                                        "3qp8j2CqCgsSa6ZKj",
+                                        "NdRNhkKWv4Xj8sxd8",
+                                        "AWLTt3qa5x6xRG353",
+                                        "cR2uFML56FL6rMf2v"
+                                    ]
+                                }
+                            ]
+                        },
+                        "results": [
+                            {
+                                "name": "predict"
+                            }
+                        ]
+                    },
+                    {
+                        "flowchartId": "2",
+                        "type": "execution",
+                        "name": "predict_property",
+                        "head": false,
+                        "execution": {
+                            "app": {
+                                "name": "ml",
+                                "exec": "rupy",
+                                "version": "0.1.0"
+                            },
+                            "preProcessors": [],
+                            "input": [
+                                {
+                                    "predict": {
+                                        "scope": "1",
+                                        "name": "predict"
+                                    },
+                                    "exabyteIds": [
+                                        "3qp8j2CqCgsSa6ZKj"
+                                    ]
+                                }
+                            ]
+                        },
+                        "results": [
+                            {
+                                "name": "band_gaps"
+                            }
+                        ]
+                    }
+                ]
+            }
+        }
+    },
+    "_project": {
+        "_id": "56c569d7fe45004bec30b5e7",
+        "name": "Project 1",
+        "owner": {
+            "_id": "sdkucJLKbsaa2ajksh",
+            "name": "demo",
+            "type": 10
+        },
+        "slug": "project-1",
+        "goldName": "exabyte-project-1"
+    },
+    "version": "0.1.0",
+    "owner": {
+        "_id": "paxJukXpDBiSRjfRZ",
+        "name": "demo",
+        "type": 10
+    },
+    "creator": {
+        "username": "demo",
+        "_id": "paxJukXpDBiSRjfRZ"
+    }
+}

example/job/theory/ml/_input_fit.json

@@ -0,0 +1,9 @@
+{
+    "arguments": {
+        "...": "include(linear/input_fit.json)"
+    },
+    "method": {
+        "type": "linear",
+        "subtype": "least_squares"
+    }
+}

example/job/theory/ml/input.json

@@ -0,0 +1,8 @@
+{
+    "...": "include(input/predict.json)",
+    "materials": {
+        "exabyteIds": [
+            "LCthJ6E2QabYCZqf3"
+        ]
+    }
+}

example/job/theory/ml/input/fit.json

@@ -0,0 +1,11 @@
+{
+    "...": "include(../_input_fit.json)",
+    "type": "fit",
+    "features": [
+        "elemental_composition",
+        "electronegativity"
+    ],
+    "properties": [
+        "band_gaps"
+    ]
+}

example/job/theory/ml/input/predict.json

@@ -0,0 +1,4 @@
+{
+    "type": "predict",
+    "method": "include(../linear.json)"
+}

example/job/theory/ml/linear.json

@@ -0,0 +1,21 @@
+{
+    "type": "linear",
+    "subtype": "least_squares",
+    "dataFrame": {
+        "...": "include(linear/data_frame.json)"
+    },
+    "data": {
+        "perProperty": [
+            {
+                "...": "include(linear/data.json)"
+            }
+        ]
+    },
+    "precision": {
+        "perProperty": [
+            {
+                "...": "include(linear/precision.json)"
+            }
+        ]
+    }
+}

example/job/theory/ml/linear/data.json

@@ -0,0 +1,27 @@
+{
+    "name": "band_gaps:indirect",
+    "coefficients": [
+        -3.339e-01,
+        2.555e-01
+    ],
+    "features": [
+        "elemental_composition",
+        "electronegativity"
+    ],
+    "intercept": [
+        0.5
+    ],
+    "scores": [
+        0.018,
+        0.513
+    ],
+    "scaler": [
+        {
+            "name": "mean",
+            "values": [
+                0.5,
+                1.95
+            ]
+        }
+    ]
+}

example/job/theory/ml/linear/data_frame.json

@@ -0,0 +1,16 @@
+{
+    "c8zG3XcnnTybfP3QQ": {
+        "electronegativity": 1.90,
+        "elemental_composition": {
+            "Ge": 0.3125,
+            "Si": 0.6875
+        }
+    },
+    "wWK2XCyivQrG8j34Z": {
+        "electronegativity": 2.00,
+        "elemental_composition": {
+            "Ge": 0.1875,
+            "Si": 0.8125
+        }
+    }
+}

example/job/theory/ml/linear/input_fit.json

@@ -0,0 +1,6 @@
+{
+    "sampleWeight": true,
+    "numFeatures": 2,
+    "scaler": "StandardScaler",
+    "numSplits":3
+}

example/job/theory/ml/linear/precision.json

@@ -0,0 +1,6 @@
+{
+    "name": "band_gaps:indirect",
+    "trainingError": 0.1,
+    "cvError": 0.3,
+    "score": 0.8
+}

example/job/workflow/unit/execution.json

@@ -5,8 +5,7 @@
     "compute": null,
     "input": [
         {
-            "content": "K_POINTS automatic\n2 2 2 0 0 0",
-            "name": "pw_scf.in"
+            "...": "include(execution/input.json)"
         }
     ]
 }

example/job/workflow/unit/execution/input.json

@@ -0,0 +1,4 @@
+{
+    "content": "K_POINTS automatic\n2 2 2 0 0 0",
+    "name": "pw_scf.in"
+}

example/material/elements/_elemental_data.json

@@ -0,0 +1,4 @@
+{
+    "name": "valence_electrons",
+    "value": 4
+}

example/material/elements/element.json

@@ -0,0 +1,41 @@
+{
+    "element": "Si",
+    "properties": [
+        {
+            "name": "valence_electrons",
+            "value": 4
+        },
+        {
+            "name": "unpaired_electrons",
+            "value": 4
+        },
+        {
+            "name": "atomic_radius",
+            "value": 111.0,
+            "units": "pm"
+        },
+        {
+            "name": "electron_configuration",
+            "value": "[Ne]3s23p2"
+        },
+        {
+            "name": "electronegativity",
+            "value": 1.90
+        },
+        {
+            "name": "ionization_potential",
+            "value": 8.15,
+            "units": "eV"
+        },
+        {
+            "name": "density",
+            "value": 2.32,
+            "units": "g/cm^3"
+        },
+        {
+            "name": "cohesive_energy",
+            "value": 4.63,
+            "units": "eV/atom"
+        }
+    ]
+}

example/material/fingerprint.json

@@ -0,0 +1,17 @@
+{
+    "elemental":[
+        {
+            "...": "include(fingerprint/elemental.json)"
+        }
+    ],
+    "primary": [
+        {
+            "...": "include(fingerprint/primary.json)"
+        }
+    ],
+    "secondary": [
+        {
+            "...": "include(characteristic.json)"
+        }
+    ]
+}

example/material/fingerprint/elemental.json

@@ -0,0 +1,6 @@
+{
+    "type": "element-specific",
+    "subtype": "numeric",
+    "name": "ionization_potential",
+    "...": "include(elemental/ionization_potential.json)"
+}

example/material/fingerprint/elemental/ionization_potential.json

@@ -0,0 +1,5 @@
+{
+    "...": "include(../schemas/element_specific.json)",
+    "name": "ionization_potential",
+    "units": "eV"
+}

example/material/fingerprint/primary.json

@@ -0,0 +1,7 @@
+{
+    "type": "compound-specific-array",
+    "subtype": "numeric",
+    "category": "stoichiometric",
+    "name": "p_norms",
+    "...": "include(primary/p_norms.json)"
+}

example/material/fingerprint/primary/p_norms.json

@@ -0,0 +1,3 @@
+{
+    "...": "include(../schemas/compound_specific_array.json)"
+}

example/material/fingerprint/schemas/compound_specific_array.json

@@ -0,0 +1,21 @@
+{
+    "name": "p_norms",
+    "data": [
+        {
+            "name": "p03",
+            "value": 0.693
+        },
+        {
+            "name": "p05",
+            "value": 0.671
+        },
+        {
+            "name": "p07",
+            "value": 0.667
+        },
+        {
+            "name": "p09",
+            "value": 0.667
+        }
+    ]
+}