csrlite 0.1.0__py3-none-any.whl → 0.2.0__py3-none-any.whl

csrlite/disposition/disposition.py

@@ -15,10 +15,10 @@ from pathlib import Path
 import polars as pl
 from rtflite import RTFDocument
 
-from ..ae.ae_utils import create_ae_rtf_table
-from ..common.count import count_subject_with_observation
+from ..common.count import count_subject, count_subject_with_observation
 from ..common.parse import StudyPlanParser
 from ..common.plan import StudyPlan
+from ..common.rtf import create_rtf_table_n_pct
 from ..common.utils import apply_common_filters
 
 
@@ -26,18 +26,21 @@ def study_plan_to_disposition_summary(
     study_plan: StudyPlan,
 ) -> list[str]:
     """
-    Generate Disposition Table 1.1 RTF outputs for all analyses defined in StudyPlan.
+    Generate Disposition Summary Table outputs for all analyses defined in StudyPlan.
     """
     # Meta data
     analysis_type = "disposition_summary"
     output_dir = study_plan.output_dir
+    title = "Disposition of Participants"
     footnote = ["Percentages are based on the number of enrolled participants."]
     source = None
 
     population_df_name = "adsl"
-    observation_df_name = "ds"  # As per plan_ds_xyz123.yaml
 
     id = ("USUBJID", "Subject ID")
+    ds_term = ("EOSSTT", "Disposition Status")
+    dist_reason_term = ("DCREASCD", "Discontinued Reason")
+
     total = True
     missing_group = "error"
 
@@ -57,32 +60,14 @@ def study_plan_to_disposition_summary(
 
     for row in disp_plans.iter_rows(named=True):
         population = row["population"]
-        observation = row.get("observation")
-        parameter = row["parameter"]
         group = row.get("group")
-        title_text = row.get(
-            "title", "Disposition of Participants"
-        )  # Allow title override from plan if we supported it in parser, else default
+        title_text = title
 
         # Get datasets
-        population_df, observation_df = parser.get_datasets(population_df_name, observation_df_name)
+        (population_df,) = parser.get_datasets(population_df_name)
 
         # Get filters
         population_filter = parser.get_population_filter(population)
-        obs_filter = parser.get_observation_filter(observation)
-
-        # Get parameters with indent levels
-        param_names, param_filters, param_labels, param_indents = parser.get_parameter_info(
-            parameter
-        )
-
-        # Apply indentation to labels
-        indented_labels = []
-        for label, indent_level in zip(param_labels, param_indents):
-            indent_str = "    " * indent_level  # 4 spaces per indent level
-            indented_labels.append(f"{indent_str}{label}")
-
-        variables_list = list(zip(param_filters, indented_labels))
 
         # Get group info (optional)
         if group is not None:
@@ -106,12 +91,11 @@ def study_plan_to_disposition_summary(
 
         rtf_path = disposition(
             population=population_df,
-            observation=observation_df,
             population_filter=population_filter,
-            observation_filter=obs_filter,
             id=id,
             group=group_tuple,
-            variables=variables_list,
+            ds_term=ds_term,
+            dist_reason_term=dist_reason_term,
             title=title_parts,
             footnote=footnote,
             source=source,
@@ -126,12 +110,11 @@ def study_plan_to_disposition_summary(
 
 def disposition(
     population: pl.DataFrame,
-    observation: pl.DataFrame,
     population_filter: str | None,
-    observation_filter: str | None,
     id: tuple[str, str],
     group: tuple[str, str] | None,
-    variables: list[tuple[str, str]],
+    ds_term: tuple[str, str],
+    dist_reason_term: tuple[str, str],
     title: list[str],
     footnote: list[str] | None,
     source: list[str] | None,
@@ -141,17 +124,16 @@ def disposition(
     missing_group: str = "error",
 ) -> str:
     """
-    Complete Disposition Table 1.1 pipeline wrapper.
+    Complete Disposition Summary Table pipeline wrapper.
     """
     # Step 1: Generate ARD
     ard = disposition_ard(
         population=population,
-        observation=observation,
         population_filter=population_filter,
-        observation_filter=observation_filter,
         id=id,
         group=group,
-        variables=variables,
+        ds_term=ds_term,
+        dist_reason_term=dist_reason_term,
         total=total,
         missing_group=missing_group,
     )
@@ -172,87 +154,135 @@ def disposition(
     return output_file
 
 
+def _validate_disposition_data(df: pl.DataFrame, ds_var: str, reason_var: str) -> None:
+    """
+    Validate disposition data integrity.
+
+    Rules:
+    1. ds_var must be {Completed, Ongoing, Discontinued} and non-null.
+    2. If ds_var is Completed/Ongoing, reason_var must be the same as ds_var or null.
+    3. If ds_var is Discontinued, reason_var must be non-null and not Completed/Ongoing.
+    """
+    # Rule 1: Valid Statuses
+    valid_statuses = ["Completed", "Ongoing", "Discontinued"]
+    if df[ds_var].is_null().any():
+        raise ValueError(f"Found null values in disposition status column '{ds_var}'")
+
+    invalid_status = df.filter(~pl.col(ds_var).is_in(valid_statuses))
+    if not invalid_status.is_empty():
+        bad_values = invalid_status[ds_var].unique().to_list()
+        raise ValueError(
+            f"Invalid disposition statuses found: {bad_values}. Must be one of {valid_statuses}"
+        )
+
+    # Rule 2: Completed/Ongoing implies Reason is Null OR equal to Status
+    inconsistent_completed = df.filter(
+        (pl.col(ds_var).is_in(["Completed", "Ongoing"]))
+        & (~pl.col(reason_var).is_null())
+        & (pl.col(reason_var) != pl.col(ds_var))
+    )
+    if not inconsistent_completed.is_empty():
+        raise ValueError(
+            f"Found subjects with status 'Completed' or 'Ongoing' with mismatched "
+            f"discontinuation reason in '{reason_var}'. Reason must be Null or match Status."
+        )
+
+    # Rule 3: Discontinued implies Reason is NOT Null AND NOT {Completed, Ongoing}
+    invalid_discontinued = df.filter(
+        (pl.col(ds_var) == "Discontinued")
+        & ((pl.col(reason_var).is_null()) | (pl.col(reason_var).is_in(["Completed", "Ongoing"])))
+    )
+    if not invalid_discontinued.is_empty():
+        raise ValueError(
+            f"Found subjects with status 'Discontinued' but missing or invalid "
+            f"discontinuation reason in '{reason_var}'"
+        )
+
+
 def disposition_ard(
     population: pl.DataFrame,
-    observation: pl.DataFrame,
     population_filter: str | None,
-    observation_filter: str | None,
     id: tuple[str, str],
     group: tuple[str, str] | None,
-    variables: list[tuple[str, str]],
+    ds_term: tuple[str, str],
+    dist_reason_term: tuple[str, str],
     total: bool,
     missing_group: str,
+    pop_var_name: str = "Enrolled",
 ) -> pl.DataFrame:
     """
-    Generate ARD for Disposition Table 1.1.
+    Generate ARD for Summary Table.
     """
+    # Unpack variables
+    ds_var_name, _ = ds_term
+    dist_reason_var_name, _ = dist_reason_term
     id_var_name, _ = id
 
-    # Handle optional group
-    if group is not None:
-        group_var_name, _ = group
-    else:
-        # Create a dummy group column for overall counts
-        group_var_name = "__all__"
-        population = population.with_columns(pl.lit("All Subjects").alias(group_var_name))
-        observation = observation.with_columns(pl.lit("All Subjects").alias(group_var_name))
-        total = False  # No need for total column when there's only one group
+    # Validate Data
+    _validate_disposition_data(population, ds_var_name, dist_reason_var_name)
 
     # Apply common filters
-    population_filtered, observation_to_filter = apply_common_filters(
+    population_filtered, _ = apply_common_filters(
         population=population,
-        observation=observation,
+        observation=None,
         population_filter=population_filter,
-        observation_filter=observation_filter,
+        observation_filter=None,
     )
 
-    # For each parameter, we create an "observation" dataset and use
-    # count_subject_with_observation. This approach works for both ADSL-based
-    # filters (e.g., "Enrolled") and DS-based filters (e.g., "Discontinued")
-
-    results = []
-
-    for var_filter, var_label in variables:
-        # Try to apply the filter to population first, then observation
-        # This handles both ADSL-based and DS-based parameter filters
-        try:
-            target_obs = population_filtered.filter(pl.sql_expr(var_filter))
-        except Exception:
-            target_obs = observation_to_filter.filter(pl.sql_expr(var_filter))
-
-        # Add the parameter label as a variable for counting
-        target_obs = target_obs.with_columns(pl.lit(var_label).alias("__index__"))
-
-        # Use count_subject_with_observation to get n (%) for each group
-        counts = count_subject_with_observation(
-            population=population_filtered,
-            observation=target_obs,
-            id=id_var_name,
-            group=group_var_name,
-            variable="__index__",
-            total=total,
-            missing_group=missing_group,
+    if group:
+        group_var_name, _ = group
+    else:
+        # Create dummy group for overall analysis
+        group_var_name = "Overall"
+        total = False
+        population_filtered = population_filtered.with_columns(
+            pl.lit("Overall").alias(group_var_name)
         )
 
-        results.append(
-            counts.select(
-                pl.col("__index__"),
-                pl.col(group_var_name).alias("__group__"),
-                pl.col("n_pct_subj_fmt").alias("__value__"),
-            )
-        )
+    # Enrolled Subjects
+    n_pop_counts = count_subject(
+        population=population_filtered,
+        id=id_var_name,
+        group=group_var_name,
+        total=total,
+        missing_group=missing_group,
+    )
+
+    n_pop = n_pop_counts.select(
+        pl.lit(pop_var_name).alias("__index__"),
+        pl.col(group_var_name).cast(pl.String).alias("__group__"),
+        pl.col("n_subj_pop").cast(pl.String).alias("__value__"),
+    )
+
+    # Hierarchical Counts for Status and Reason
+    # Level 1: Status (Completed, Ongoing, Discontinued)
+    # Level 2: Status + Reason (Only relevant for Discontinued)
+    n_dict = count_subject_with_observation(
+        population=population_filtered,
+        observation=population_filtered,
+        id=id_var_name,
+        group=group_var_name,
+        variable=[ds_var_name, dist_reason_var_name],
+        total=total,
+        missing_group=missing_group,
+    )
+
+    # Filter and format
+    # Identify rows:
+    # 1. Status rows: Where reason is "__all__"
+    # 2. Reason rows: Where reason is specific value (indented)
+    n_dict = n_dict.unique([group_var_name, ds_var_name, dist_reason_var_name, "__id__"])
 
-    # Combine all results
-    ard = pl.concat(results)
+    # Filter out redundant nested rows (e.g., "Completed" under "Completed")
+    n_dict = n_dict.filter(pl.col(dist_reason_var_name) != pl.col(ds_var_name))
 
-    # Sort by the order of variables in the list
-    # Create an Enum for __index__
-    var_labels = [label for _, label in variables]
-    ard = ard.with_columns(pl.col("__index__").cast(pl.Enum(var_labels))).sort(
-        "__index__", "__group__"
+    n_final = n_dict.sort("__id__").select(
+        pl.col("__variable__").alias("__index__"),
+        pl.col(group_var_name).cast(pl.String).alias("__group__"),
+        pl.col("n_pct_subj_fmt").cast(pl.String).alias("__value__"),
     )
 
-    return ard
+    return pl.concat([n_pop, n_final])
 
 
 def disposition_df(ard: pl.DataFrame) -> pl.DataFrame:
@@ -260,10 +290,11 @@ def disposition_df(ard: pl.DataFrame) -> pl.DataFrame:
     Transform ARD to display format.
     """
     # Pivot
+    # Pivot from long to wide format
     df_wide = ard.pivot(index="__index__", on="__group__", values="__value__")
 
-    # Rename index
-    df_wide = df_wide.rename({"__index__": "Disposition Status"})
+    # Rename __index__ to display column name
+    df_wide = df_wide.rename({"__index__": "Term"}).select(pl.col("Term"), pl.exclude("Term"))
 
     return df_wide
 
@@ -282,7 +313,7 @@ def disposition_rtf(
     # Columns: Disposition Status, Group 1, Group 2, ... Total
 
     n_cols = len(df.columns)
-    col_header_1 = list(df.columns)
+    col_header_1 = [""] + list(df.columns[1:])
     col_header_2 = [""] + ["n (%)"] * (n_cols - 1)
 
     if col_rel_width is None:
@@ -290,7 +321,7 @@ def disposition_rtf(
     else:
         col_widths = col_rel_width
 
-    return create_ae_rtf_table(
+    return create_rtf_table_n_pct(
         df=df,
         col_header_1=col_header_1,
         col_header_2=col_header_2,

{csrlite-0.1.0.dist-info → csrlite-0.2.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: csrlite
-Version: 0.1.0
+Version: 0.2.0
 Summary: A hierarchical YAML-based framework for generating Tables, Listings, and Figures in clinical trials
 Author-email: Clinical Biostatistics Team <biostat@example.com>
 License: MIT
@@ -28,17 +28,17 @@ Provides-Extra: plotting
 Requires-Dist: matplotlib>=3.5.0; extra == "plotting"
 Requires-Dist: plotly>=5.0.0; extra == "plotting"
 Provides-Extra: dev
-Requires-Dist: pytest>=7.0.0; extra == "dev"
 Requires-Dist: pytest-cov>=4.0.0; extra == "dev"
-Requires-Dist: black>=22.0.0; extra == "dev"
-Requires-Dist: isort>=5.0.0; extra == "dev"
-Requires-Dist: mypy>=1.0.0; extra == "dev"
 Requires-Dist: pytest>=9.0.1; extra == "dev"
+Requires-Dist: black>=22.0.0; extra == "dev"
+Requires-Dist: isort>=7.0.0; extra == "dev"
+Requires-Dist: ruff>=0.14.8; extra == "dev"
+Requires-Dist: mypy>=1.19.0; extra == "dev"
+Requires-Dist: quarto>=0.1.0; extra == "dev"
+Requires-Dist: pyre-check>=0.9.18; extra == "dev"
 Requires-Dist: jupyter>=1.1.1; extra == "dev"
 Requires-Dist: jupyter-cache>=1.0.1; extra == "dev"
 Requires-Dist: nbformat>=5.10.4; extra == "dev"
-Requires-Dist: ruff>=0.1.0; extra == "dev"
-Requires-Dist: pyre-check>=0.9.18; extra == "dev"
 Provides-Extra: all
 Requires-Dist: rtflite; extra == "all"
 Requires-Dist: matplotlib>=3.5.0; extra == "all"

csrlite-0.2.0.dist-info/RECORD

@@ -0,0 +1,19 @@
+csrlite/__init__.py,sha256=w18H3dEZE_HZFdsr0Qqo0xmJPksx764BFAg8XPo9tGI,1417
+csrlite/ae/__init__.py,sha256=gZHPLATRF9f8QBwwQtEjQRtXMsqOJsUK2sbUMLjiE5U,14
+csrlite/ae/ae_listing.py,sha256=R4g8JnJRCx4u60xRC2IEu8EPzEIgxj8VgDhBaaQ5eZE,18389
+csrlite/ae/ae_specific.py,sha256=_CDAgF4vMmjpqFTppL2LgmExVsEhA57z4jYj4Y1zLfY,17022
+csrlite/ae/ae_summary.py,sha256=46IyuqHGdn0dLOrz7XffKNcNjscA0Y8OZFiZ6akisB4,13692
+csrlite/ae/ae_utils.py,sha256=ew5Mm_zNdflc_MRYvYSChXhRhGQ1oZcz7H_TZPVvFBk,2011
+csrlite/common/config.py,sha256=FUnUL1BtQO52U0ag1U_d2K3UP5L_vA_KifonANHLv_c,878
+csrlite/common/count.py,sha256=k1W-LdQv63s-B-Oeq2SvYsXctrT1YMVWs93CtIaGpVw,8785
+csrlite/common/parse.py,sha256=Vz9C7ljkDygT2qkP6TlY3T3p71D6BD5GtIwRKv6p8ps,9319
+csrlite/common/plan.py,sha256=XXUGpzNxC6oS66c7NYnDPmE0CwXMhIQzlJCga1nDktw,12928
+csrlite/common/rtf.py,sha256=gah-M-WdvMk52R-AEacM79P18jc2OFnCH7-I0B91Fhk,2825
+csrlite/common/utils.py,sha256=It0aHqPfXDmCte2uVAO2Lkb3U_jDLrjNihAL8gziTQk,1110
+csrlite/common/yaml_loader.py,sha256=_v9pkbAUVshTqVoMLqMiEn17awL2K0kFR4pdDArMSOM,3071
+csrlite/disposition/__init__.py,sha256=KMtGoBjN4aKNYvXHmZ0GX-f4RnmQ3coYbUrkFeU8Es0,85
+csrlite/disposition/disposition.py,sha256=r0R53ozVFYNXI0JBzR14sghQayLVUvXJBiS0wNLAtB0,10348
+csrlite-0.2.0.dist-info/METADATA,sha256=SzOst7_7EmNn6FB29KO-oFZpGZ93yZCQ9lGBhG-p9bE,2801
+csrlite-0.2.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+csrlite-0.2.0.dist-info/top_level.txt,sha256=59zJTvGH5zx2FY4vCl4kgnH8awT0cZrg21Mace7IFlU,8
+csrlite-0.2.0.dist-info/RECORD,,

csrlite-0.1.0.dist-info/RECORD

@@ -1,17 +0,0 @@
-csrlite/__init__.py,sha256=o7HOFA9KKbyfq8l_26dqNHBDz2jqDJm8lQBvXYfBYdQ,1164
-csrlite/ae/__init__.py,sha256=gZHPLATRF9f8QBwwQtEjQRtXMsqOJsUK2sbUMLjiE5U,14
-csrlite/ae/ae_listing.py,sha256=EwmU5CTmqmkuiOsA7FedEF83S9MJ1YPlmf5AMsksUCU,18343
-csrlite/ae/ae_specific.py,sha256=s-Zj6WQhKKHARMt5LkoSw74iThPVd-_92l8eYhUGiPc,16898
-csrlite/ae/ae_summary.py,sha256=NlqbuW0N0aiJ6i3fLCDJJPxPuqk1mv6i5svPsIT1xD0,13637
-csrlite/ae/ae_utils.py,sha256=6UhUrTkyOgpxpl5YFoNjteLBgkf0Gtw5lgQApCkwf3c,4121
-csrlite/common/count.py,sha256=gdTSlA-nr5B6e3fuP9pelASf_FdaeRKYzujpE0bbzvA,6925
-csrlite/common/parse.py,sha256=Vz9C7ljkDygT2qkP6TlY3T3p71D6BD5GtIwRKv6p8ps,9319
-csrlite/common/plan.py,sha256=QhsBD7b-AU_mc-JScLHM1Oiw7FJ4AKN1iHWX80-ukuw,11988
-csrlite/common/utils.py,sha256=SAqEnwDtE32LuQqnMVQr_1Xfdp-z54wIrwbwwPBE9lU,1022
-csrlite/common/yaml_loader.py,sha256=_v9pkbAUVshTqVoMLqMiEn17awL2K0kFR4pdDArMSOM,3071
-csrlite/disposition/__init__.py,sha256=KMtGoBjN4aKNYvXHmZ0GX-f4RnmQ3coYbUrkFeU8Es0,85
-csrlite/disposition/disposition.py,sha256=UMm4Z1fFQ6VJ-KSqSaMP7qEzLoSa399kRSZx-oPKEqM,9274
-csrlite-0.1.0.dist-info/METADATA,sha256=HNbQnMH5LqJ2Vq_KMky8u9_ofP0YOe40AynZM9384FI,2799
-csrlite-0.1.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-csrlite-0.1.0.dist-info/top_level.txt,sha256=59zJTvGH5zx2FY4vCl4kgnH8awT0cZrg21Mace7IFlU,8
-csrlite-0.1.0.dist-info/RECORD,,