Spaces:
Sleeping
Sleeping
Upload openenv.yaml with huggingface_hub
Browse files- openenv.yaml +98 -0
openenv.yaml
ADDED
|
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: data-cleaning-openenv
|
| 2 |
+
version: 1.0.0
|
| 3 |
+
description: >
|
| 4 |
+
An OpenEnv-compliant environment where AI agents learn to clean
|
| 5 |
+
messy real-world datasets step by step. Tasks include duplicate
|
| 6 |
+
removal, missing value imputation, dtype fixing, outlier detection,
|
| 7 |
+
and schema validation.
|
| 8 |
+
|
| 9 |
+
author: thorodin103
|
| 10 |
+
tags:
|
| 11 |
+
- openenv
|
| 12 |
+
- data-cleaning
|
| 13 |
+
- real-world
|
| 14 |
+
- pandas
|
| 15 |
+
- tabular
|
| 16 |
+
|
| 17 |
+
tasks:
|
| 18 |
+
- id: easy_dedup_rename
|
| 19 |
+
difficulty: easy
|
| 20 |
+
description: Remove duplicate rows and rename columns to snake_case
|
| 21 |
+
max_steps: 10
|
| 22 |
+
scoring:
|
| 23 |
+
duplicate_score: 0.5
|
| 24 |
+
schema_score: 0.5
|
| 25 |
+
|
| 26 |
+
- id: medium_missing_dtype
|
| 27 |
+
difficulty: medium
|
| 28 |
+
description: Fill missing values and fix wrong data types
|
| 29 |
+
max_steps: 15
|
| 30 |
+
scoring:
|
| 31 |
+
missing_score: 0.5
|
| 32 |
+
dtype_score: 0.5
|
| 33 |
+
|
| 34 |
+
- id: hard_full_pipeline
|
| 35 |
+
difficulty: hard
|
| 36 |
+
description: Full pipeline — duplicates, missing, dtypes, outliers, schema
|
| 37 |
+
max_steps: 20
|
| 38 |
+
scoring:
|
| 39 |
+
duplicate_score: 0.2
|
| 40 |
+
missing_score: 0.2
|
| 41 |
+
dtype_score: 0.2
|
| 42 |
+
outlier_score: 0.2
|
| 43 |
+
schema_score: 0.2
|
| 44 |
+
|
| 45 |
+
observation_space:
|
| 46 |
+
task_id: string
|
| 47 |
+
step: integer
|
| 48 |
+
dataset_info: object
|
| 49 |
+
columns: list
|
| 50 |
+
shape: list
|
| 51 |
+
missing_values: object
|
| 52 |
+
dtypes: object
|
| 53 |
+
duplicate_count: integer
|
| 54 |
+
sample_rows: list
|
| 55 |
+
available_operations: list
|
| 56 |
+
task_description: string
|
| 57 |
+
message: string
|
| 58 |
+
|
| 59 |
+
action_space:
|
| 60 |
+
operation:
|
| 61 |
+
type: string
|
| 62 |
+
values:
|
| 63 |
+
- remove_duplicates
|
| 64 |
+
- fill_missing
|
| 65 |
+
- fix_dtype
|
| 66 |
+
- remove_outliers
|
| 67 |
+
- rename_columns
|
| 68 |
+
- validate_schema
|
| 69 |
+
- finish
|
| 70 |
+
parameters:
|
| 71 |
+
type: object
|
| 72 |
+
description: Operation specific parameters
|
| 73 |
+
|
| 74 |
+
reward_space:
|
| 75 |
+
total:
|
| 76 |
+
type: float
|
| 77 |
+
range: [0.0, 1.0]
|
| 78 |
+
components:
|
| 79 |
+
- duplicate_score
|
| 80 |
+
- missing_score
|
| 81 |
+
- dtype_score
|
| 82 |
+
- outlier_score
|
| 83 |
+
- schema_score
|
| 84 |
+
- penalty
|
| 85 |
+
|
| 86 |
+
api:
|
| 87 |
+
reset: POST /reset/{task_id}
|
| 88 |
+
step: POST /step/{task_id}
|
| 89 |
+
state: GET /state/{task_id}
|
| 90 |
+
tasks: GET /tasks
|
| 91 |
+
validate: GET /validate
|
| 92 |
+
health: GET /health
|
| 93 |
+
|
| 94 |
+
runtime:
|
| 95 |
+
language: python
|
| 96 |
+
version: "3.10"
|
| 97 |
+
framework: fastapi
|
| 98 |
+
port: 7860
|