Remove old demos, fix airbnb_reviews and add store demo (#131)

trane-dev · Aug 10, 2023 · 2cfe08b · 2cfe08b
1 parent 093b1e4
commit 2cfe08b
Show file tree

Hide file tree

Showing 24 changed files with 370,546 additions and 220,210 deletions.
diff --git a/.github/workflows/test_without_dev_deps.yaml b/.github/workflows/test_without_dev_deps.yaml
@@ -37,26 +37,25 @@ jobs:
       - name: Run simple usage
         run: |
           import trane
-          data = trane.datasets.load_covid()
-          table_meta = trane.datasets.load_covid_metadata()
-          entity_col = "Country/Region"
-          window_size = "2d"
-          minimum_data = "2020-01-22"
-          maximum_data = "2020-03-29"
+          df = trane.datasets.load_airbnb_reviews()
+          entity_col = "location"
+          time_col = "date"
+          window_size = "1m"
+          table_meta = trane.typing.infer_table_meta(df, entity_col, time_col)
+          table_meta["location"] = trane.typing.column_schema.ColumnSchema(
+              trane.typing.logical_types.Categorical, semantic_tags={"category", "primary_key"}
+          )
           cutoff_strategy = trane.CutoffStrategy(
               entity_col=entity_col,
               window_size=window_size,
-              minimum_data=minimum_data,
-              maximum_data=maximum_data,
           )
-          time_col = "Date"
           problem_generator = trane.PredictionProblemGenerator(
-              df=data,
+              df=df,
               entity_col=entity_col,
               time_col=time_col,
               cutoff_strategy=cutoff_strategy,
               table_meta=table_meta,
           )
-          problems = problem_generator.generate(data, generate_thresholds=True)
-          assert len(problems) >= 800
+          problems = problem_generator.generate(df, generate_thresholds=True)
+          assert len(problems) >= 450
         shell: python
diff --git a/docs/changelog.md b/docs/changelog.md
@@ -8,16 +8,19 @@ v0.6.0 (, 2023)
 * Enhancements
     * Add pyarrow dependency and use pyarrow backed dtypes [#120][#120]
     * Add Airbnb Reviews dataset [#125][#125]
+    * Add Store dataset [#131][#131]
 * Fixes
     * Rename `_execute_operations_on_df` to `target` in executed prediction problem dataframe [#124][#124]
     * Clean up operation description generation [#118][#118]
     * Remove PredictionProblemEvaluator [#118][#118]
     * Remove FeaturetoolsWrapper class [#100][#100]
+    * Remove covid19 and youtube datasets [#131][#131]
 
     [#124]: <https://github.com/trane-dev/Trane/pull/124>
     [#118]: <https://github.com/trane-dev/Trane/pull/118>
     [#120]: <https://github.com/trane-dev/Trane/pull/120>
     [#125]: <https://github.com/trane-dev/Trane/pull/125>
+    [#131]: <https://github.com/trane-dev/Trane/pull/131>
 
 
 v0.5.0 (July 27, 2023)

diff --git a/pyproject.toml b/pyproject.toml
@@ -66,7 +66,7 @@ dev = [
 
 [tool.setuptools]
 include-package-data = true
-license-files = ["LICENSE"]
+license-files = ["LICENSE", "trane/datasets/data/store/LICENSE"]
 
 [tool.setuptools.packages.find]
 namespaces = true
@@ -79,8 +79,15 @@ namespaces = true
     "Makefile",
 ]
 "trane" = [
-    "datasets/data/covid19.csv",
-    "datasets/data/USvideos.csv",
+    "trane/datasets/data/airbnb_reviews/airbnb_reviews.csv",
+    "trane/datasets/data/store/categories.csv",
+    "trane/datasets/data/store/cust_hist.csv",
+    "trane/datasets/data/store/customers.csv",
+    "trane/datasets/data/store/inventory.csv",
+    "trane/datasets/data/store/orderlines.csv",
+    "trane/datasets/data/store/orders.csv",
+    "trane/datasets/data/store/products.csv",
+    "trane/datasets/data/store/reorder.csv",
 ]
 
 [tool.setuptools.exclude-package-data]