Mistral-prompter

sigllm/pipelines/prompter/mistral_prompter.json

@@ -0,0 +1,66 @@
+{
+    "primitives": [
+        "mlstars.custom.timeseries_preprocessing.time_segments_aggregate",
+        "sklearn.impute.SimpleImputer",
+        "sigllm.primitives.transformation.Float2Scalar",
+        "sigllm.primitives.prompting.timeseries_preprocessing.rolling_window_sequences",
+	    "sigllm.primitives.transformation.format_as_string",
+        "sigllm.primitives.prompting.huggingface.HF",
+        "sigllm.primitives.transformation.format_as_integer",
+        "sigllm.primitives.prompting.anomalies.val2idx",
+        "sigllm.primitives.prompting.anomalies.ano_within_windows",
+        "sigllm.primitives.prompting.anomalies.merge_anomaly_seq",
+        "sigllm.primitives.prompting.anomalies.idx2time",
+        "sigllm.primitives.prompting.anomalies.timestamp2interval"
+    ],
+    "init_params": {
+        "mlstars.custom.timeseries_preprocessing.time_segments_aggregate#1": {
+            "time_column": "timestamp",
+            "interval": 21600,
+            "method": "mean"
+        },
+        "sigllm.primitives.transformation.Float2Scalar#1": {
+            "decimal": 2,
+            "rescale": true
+        },
+        "sigllm.primitives.prompting.timeseries_preprocessing.rolling_window_sequences#1": {
+            "window_size": 200,
+            "step_size": 40
+        },
+        "sigllm.primitives.transformation.format_as_string#1": {
+            "space": false
+        },
+        "sigllm.primitives.prompting.huggingface.HF#1": {
+            "name": "mistralai/Mistral-7B-Instruct-v0.2",
+            "samples": 10
+        },
+        "sigllm.primitives.prompting.anomalies.ano_within_windows#1": {
+            "alpha": 0.4
+        },
+        "sigllm.primitives.prompting.anomalies.merge_anomaly_seq#1": {
+            "beta": 0.5
+        }
+    },
+    "input_names": {
+    "sigllm.primitives.prompting.huggingface.HF#1": {
+        "X": "X_str"
+    },
+    "sigllm.primitives.transformation.format_as_integer#1":{
+        "X": "y_hat"
+    }
+    },
+    "output_names": {
+    "mlstars.custom.timeseries_preprocessing.time_segments_aggregate#1": {
+        "index": "timestamp"
+    },
+    "sigllm.primitives.transformation.format_as_string#1": {
+        "X": "X_str"
+    },
+    "sigllm.primitives.prompting.huggingface.HF#1": {
+        "y": "y_hat"
+    },
+    "sigllm.primitives.transformation.format_as_integer#1":{
+        "X": "y"
+    }
+    }
+}

sigllm/primitives/jsons/sigllm.primitives.prompting.anomalies.ano_within_windows.json

@@ -0,0 +1,36 @@
+{
+    "name": "sigllm.primitives.prompting.anomalies.ano_within_windows",
+    "contributors": [
+        "Sarah Alnegheimish <[email protected]>",
+        "Linh Nguyen <[email protected]>"
+    ],
+    "description": "Get the final list of anomalous indices of each window",
+    "classifiers": {
+        "type": "postprocessor",
+        "subtype": "merger"
+    },
+    "modalities": [],
+    "primitive": "sigllm.primitives.prompting.anomalies.ano_within_windows",
+    "produce": {
+        "args": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            }
+        ],
+        "output": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            }
+        ]
+    }, 
+    "hyperparameters": {
+        "fixed": {
+            "alpha": {
+                "type": "float",
+                "default": 0.5
+            }
+        }
+    }
+}

sigllm/primitives/jsons/sigllm.primitives.prompting.anomalies.idx2time.json

@@ -0,0 +1,32 @@
+{
+    "name": "sigllm.primitives.prompting.anomalies.idx2time",
+    "contributors": [
+        "Sarah Alnegheimish <[email protected]>",
+        "Linh Nguyen <[email protected]>"
+    ],
+    "description": "Convert list of indices into list of timestamp",
+    "classifiers": {
+        "type": "postprocessor",
+        "subtype": "converter"
+    },
+    "modalities": [],
+    "primitive": "sigllm.primitives.prompting.anomalies.idx2time",
+    "produce": {
+        "args": [
+            {
+                "name": "timestamp",
+                "type": "ndarray"
+            }, 
+            {
+                "name": "y",
+                "type": "ndarray"
+            }
+        ],
+        "output": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            }
+        ]
+    }
+}

sigllm/primitives/jsons/sigllm.primitives.prompting.anomalies.merge_anomaly_seq.json

@@ -0,0 +1,48 @@
+{
+    "name": "sigllm.primitives.prompting.anomalies.merge_anomaly_seq",
+    "contributors": [
+        "Sarah Alnegheimish <[email protected]>",
+        "Linh Nguyen <[email protected]>"
+    ],
+    "description": "Get the final list of anomalous indices of a sequence when merging all rolling windows",
+    "classifiers": {
+        "type": "postprocessor",
+        "subtype": "merger"
+    },
+    "modalities": [],
+    "primitive": "sigllm.primitives.prompting.anomalies.merge_anomaly_seq",
+    "produce": {
+        "args": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            }, 
+            {
+                "name": "first_index",
+                "type": "ndarray"
+            },
+            {
+                "name": "window_size",
+                "type": "int"
+            },
+            {
+                "name": "step_size",
+                "type": "int"
+            }
+        ],
+        "output": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            }
+        ]
+    }, 
+    "hyperparameters": {
+        "fixed": {
+            "beta": {
+                "type": "float",
+                "default": 0.5
+            }
+        }
+    }
+}

sigllm/primitives/jsons/sigllm.primitives.prompting.anomalies.timestamp2interval.json

@@ -0,0 +1,40 @@
+{
+    "name": "sigllm.primitives.prompting.anomalies.timestamp2interval",
+    "contributors": [
+        "Sarah Alnegheimish <[email protected]>",
+        "Linh Nguyen <[email protected]>"
+    ],
+    "description": "Convert list of timestamps to list of intervals by padding to both sides and merge overlapping",
+    "classifiers": {
+        "type": "postprocessor",
+        "subtype": "converter"
+    },
+    "modalities": [],
+    "primitive": "sigllm.primitives.prompting.anomalies.timestamp2interval",
+    "produce": {
+        "args": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            },
+            {
+                "name": "timestamp",
+                "type": "ndarray"
+            }
+        ],
+        "output": [
+            {
+                "name": "df",
+                "type": "Dataframe"
+            }
+        ]
+    }, 
+    "hyperparameters": {
+        "fixed": {
+            "padding_size": {
+                "type": "int",
+                "default": 50
+            }
+        }
+    }
+}

sigllm/primitives/jsons/sigllm.primitives.prompting.anomalies.val2idx.json

@@ -0,0 +1,32 @@
+{
+    "name": "sigllm.primitives.prompting.anomalies.val2idx",
+    "contributors": [
+        "Sarah Alnegheimish <[email protected]>",
+        "Linh Nguyen <[email protected]>"
+    ],
+    "description": "Convert detected anomalous values into indices",
+    "classifiers": {
+        "type": "postprocessor",
+        "subtype": "converter"
+    },
+    "modalities": [],
+    "primitive": "sigllm.primitives.prompting.anomalies.val2idx",
+    "produce": {
+        "args": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            },
+            {
+                "name": "X",
+                "type": "ndarray"
+            }
+        ],
+        "output": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            }
+        ]
+    }
+}

sigllm/primitives/jsons/sigllm.primitives.prompting.gpt.GPT.json

@@ -0,0 +1,73 @@
+{
+    "name": "sigllm.primitives.prompting.gpt.GPT",
+    "contributors": [
+        "Linh Nguyen <[email protected]>"
+    ],
+    "description": "Prompt openai GPT model to detect time series anomalies.",
+    "classifiers": {
+        "type": "estimator",
+        "subtype": "detector"
+    },
+    "modalities": [],
+    "primitive": "sigllm.primitives.prompting.huggingface.HF",
+    "produce": {
+        "method": "detect",
+        "args": [
+            {
+                "name": "X",
+                "type": "ndarray"
+            }
+        ],
+        "output": [
+            {
+                "name": "y",
+                "type": "ndarray"
+            },
+            {
+                "name": "logprob",
+                "type": "ndarray",
+                "default": null
+            }
+        ]
+    },
+    "hyperparameters": {
+        "fixed": {
+            "name": {
+                "type": "str",
+                "default": "gpt-3.5-turbo"
+            },
+            "sep": {
+                "type": "str",
+                "default": ","
+            },
+            "anomalous_percent": {
+                "type": "float",
+                "default": "0.5"
+            },
+            "temp": {
+                "type": "float",
+                "default": 1
+            },
+            "top_p": {
+                "type": "float",
+                "default": 1
+            },
+            "logprobs": {
+                "type": "bool",
+                "default": false
+            },
+            "top_logprobs": {
+                "type": "int",
+                "default": null
+            },
+            "samples": {
+                "type": "int",
+                "default": 1
+            },
+            "seed": {
+                "type": "int",
+                "default": null
+            }
+        }
+    }
+}