From 93eb7d9efb7fe2719a3b95919585b2ff4bb9c98c Mon Sep 17 00:00:00 2001
From: jvivian <jtvivian@gmail.com>
Date: Thu, 13 Jun 2024 19:18:10 -0700
Subject: [PATCH] Rename repository before making public Fixes #73

---
 .../streamlit/pages/1_Factor_Analysis.py      | 87 +++++++++++++------
 1 file changed, 59 insertions(+), 28 deletions(-)

diff --git a/covid19_drdfm/streamlit/pages/1_Factor_Analysis.py b/covid19_drdfm/streamlit/pages/1_Factor_Analysis.py
index 1016899..5f9becf 100644
--- a/covid19_drdfm/streamlit/pages/1_Factor_Analysis.py
+++ b/covid19_drdfm/streamlit/pages/1_Factor_Analysis.py
@@ -36,28 +36,26 @@ def normalize(df):
 # TEST_DIR = Path('covid19_drdfm/data/example-output/')
 
 # Parameter for results
+def get_factors(res_dir):
+    factor_path = res_dir / "factors.csv"
+    df = pd.read_csv(factor_path, index_col=0)
+    df["Time"] = df.index
+    df.index.name = "Time"
+    cols_to_drop = [x for x in df.columns if "Time." in x]
+    df = df.drop(columns=cols_to_drop)
+    df.columns = [x.lstrip("Factor_") for x in df.columns]
+    return df
+
+
 res_dir = Path(st.text_input("Path to results", value=EX_PATH))
-factor_path = res_dir / "factors.csv"
-df = pd.read_csv(factor_path, index_col=0)
-df["Time"] = df.index
-df.index.name = "Time"
-cols_to_drop = [x for x in df.columns if "Time." in x]
-df = df.drop(columns=cols_to_drop)
-df.columns = [x.lstrip("Factor_") for x in df.columns]
-# st.dataframe(df)
-
-# AnnData
-# dfs = []
-# for subdir in res_dir.iterdir():
-#     if not subdir.is_dir():
-#         continue
-#     dfs.append(pd.read_csv(res_dir / 'df.csv'))
+if not res_dir:
+    st.warning("Please provide and hit <ENTER>")
+    st.stop()
+df = get_factors(res_dir)
 
 filter_list = ["Unnamed", "Time", "State"]
-factor = st.sidebar.selectbox("Factor", [x for x in df.columns if x not in filter_list])
 state = st.sidebar.selectbox("State", sorted(df.State.unique()))
 
-
 with st.expander("State Factors"):
     st.dataframe(df[df.State == state])
 
@@ -70,28 +68,61 @@ def normalize(df):
 # factor_vars = [x for x in FACTORS_GROUPED[factor.split("_")[1]] if x in valid_cols]  # and '_' in x]
 # columns = [*factor_vars, "State", "Time"]
 
+# Make df from res_dir
+dfs = []
+for subdir in res_dir.iterdir():
+    if not subdir.is_dir():
+        continue
+    path = subdir / "df.csv"
+    if not path.exists():
+        st.write(f"Skipping {path}, not found")
+        continue
+    sub = pd.read_csv(path, index_col=0)
+    sub["State"] = subdir.stem
+    dfs.append(sub)
+
+new = pd.concat(dfs)
 
 # Normalize original data for state / valid variables
 ad = ann.read_h5ad(res_dir / "data.h5ad")
-new = ad.to_df().reset_index()
-new["State"] = ad.obs["State"].to_list()
-new = normalize(new[new.State == state])
+factor_map = ad.var["factor"].to_frame()
+factor_set = factor_map["factor"].unique().to_list() + [x for x in df.columns if "Global" in x]
+# st.dataframe(factor_map)
+factor = st.sidebar.selectbox("Factor", factor_set)
+# new = ad.to_df().reset_index()
+# new["State"] = ad.obs["State"].to_list()
+# new = normalize(new[new.State == state])
 
 # Normalize factors and add to new dataframe
-if st.sidebar.checkbox("Invert Factor"):
-    df[factor] = df[factor] * -1
-df = normalize(df[df.State == state]).reset_index(drop=True)
+# if st.sidebar.checkbox("Invert Factor"):
+# df[factor] = df[factor] * -1
+# df = normalize(df[df.State == state]).reset_index(drop=True)
+
+df = df[df["State"] == state]
+df = df[[factor]].join(new, on="Time")
 
+# st.dataframe(df.head())
+# st.dataframe(new.head())
 
 # Coerce time bullshit to get dates standardized
-df["Time"] = pd.to_datetime(df["Time"]).dt.date
-new["Time"] = pd.to_datetime(new["Time"]).dt.date
-df = df[[factor, "Time"]].merge(new, on="Time")
+# df["Time"] = pd.to_datetime(df["Time"]).dt.date
+# new["Time"] = pd.to_datetime(new["Time"]).dt.date
 with st.expander("Graph Data"):
-    st.dataframe(df)
+    factor_cols = factor_map[factor_map["factor"] == factor]
+    if factor_cols.empty:
+        factor_cols = new.columns
+    else:
+        factor_cols = factor_cols.index.to_list()
+        factor_cols += [factor]
+    factor_cols = [x for x in factor_cols if x in df.columns]
+    st.write(factor_cols)
+    st.dataframe(df[factor_cols])
+
+df = df[factor_cols].reset_index()
 
 # Melt into format for plotting
-melted_df = df.drop(columns="State").melt(id_vars=["Time"], value_name="value")
+# melted_df = df.drop(columns="State").melt(id_vars=["Time"], value_name="value")
+melted_df = df.melt(id_vars=["Time"], value_name="value")
 melted_df["Label"] = [5 if x == factor else 1 for x in melted_df.variable]
 
 # Plot