Merge pull request #26 from viventriglia/develop

v0.2.1

src/aggregation.py

@@ -114,10 +114,8 @@ def get_portfolio_pivot(
         .sum()
         .reset_index()
     )
-    df_pivot["weight_pf"] = (
-        (100 * df_pivot["position_value"].div(pf_actual_value)).astype(float).round(1)
-    )
-    df_pivot["position_value"] = df_pivot["position_value"].astype(float).round(1)
+    df_pivot["weight_pf"] = 100 * df_pivot["position_value"].div(pf_actual_value)
+    df_pivot["position_value"] = df_pivot["position_value"]
     return df_pivot
 
 
@@ -128,8 +126,6 @@ def get_pnl_by_asset_class(
     group_by: Literal["asset_class", "macro_asset_class"],
 ) -> pd.DataFrame:
     df_pnl = df.merge(df_dimensions, how="left", on="ticker_yf")
-    df_pnl["pnl"] = np.round(
-        ((df_pnl["price"] - df_pnl["dca"]) * df_pnl["shares"]).astype(float), 1
-    )
+    df_pnl["pnl"] = ((df_pnl["price"] - df_pnl["dca"]) * df_pnl["shares"]).astype(float)
     df_pnl = df_pnl.groupby(group_by)["pnl"].sum().reset_index().sort_values([group_by])
     return df_pnl

src/pages/1_🎯_Asset_Allocation_&_PnL.py

@@ -51,18 +51,15 @@
 consider_fees = st.checkbox("Take fees into account")
 
 pf_actual_value = (df_j["shares"] * df_j["price"]).sum()
-if consider_fees:
-    pnl = pf_actual_value - expense - fees
-    pnl_perc = 100 * (pf_actual_value - expense - fees) / expense
-else:
-    pnl = pf_actual_value - expense
-    pnl_perc = 100 * (pf_actual_value - expense) / expense
+total_expense = expense + fees if consider_fees else expense
+pnl = pf_actual_value - total_expense
+pnl_perc = 100 * pnl / total_expense
 sign = "+" if pnl >= 0 else ""
 
 col_l.metric(
     label="Actual Portfolio Value",
-    value=f"{pf_actual_value: .1f} €",
-    delta=f"{pnl: .1f} € ({sign}{pnl_perc: .1f}%)",
+    value=f"{pf_actual_value: ,.1f} €",
+    delta=f"{pnl: ,.1f} € ({sign}{pnl_perc: .1f}%)",
 )
 
 df_j["position_value"] = df_j["shares"] * df_j["price"]
@@ -83,6 +80,7 @@
     group_by = st.radio(
         label="Aggregate by:",
         options=["Macro Asset Classes", "Asset Classes", "Tickers"],
+        index=1,
         horizontal=True,
     )
     df_pivot_ = get_portfolio_pivot(
@@ -91,18 +89,24 @@
         pf_actual_value=pf_actual_value,
         aggregation_level=DICT_GROUPBY_LEVELS[group_by],
     )
-    df_pivot_.index += 1
-    st.table(
+    st.dataframe(
         df_pivot_.rename(
             columns={
                 "macro_asset_class": "Macro Asset Class",
                 "asset_class": "Asset Class",
                 "ticker_yf": "Ticker",
                 "name": "Name",
-                "position_value": "Position Value (€)",
-                "weight_pf": "Weight (%)",
+                "position_value": "Position Value",
+                "weight_pf": "Weight",
+            }
+        ).style.format(
+            {
+                "Position Value": "{:,.1f} €",
+                "Weight": "{:,.1f}%",
             }
-        ).style.format(precision=1)
+        ),
+        use_container_width=True,
+        hide_index=True,
     )
 
 st.markdown("***")

src/pages/2_📈_Return_Analysis.py

@@ -28,7 +28,12 @@
     st.error("Oops... there's nothing to display. Go through 🏠 first to load the data")
     st.stop()
 
-ticker_list = df_transactions["ticker_yf"].unique().tolist()
+df_n_shares = (
+    df_transactions.groupby("ticker_yf")
+    .agg(n_shares=("shares", "sum"))
+    .sort_values("n_shares", ascending=False)
+)
+ticker_list = df_n_shares.loc[~(df_n_shares == 0).all(axis=1)].index.unique().to_list()
 df_common_history = get_max_common_history(ticker_list=ticker_list)
 
 st.markdown("## Global settings")
@@ -79,6 +84,7 @@
 df_rets = get_period_returns(
     df=df_common_history.loc[first_day:last_day, :],
     df_registry=df_registry,
+    tickers_to_evaluate=ticker_list,
     period=DICT_FREQ_RESAMPLE[freq],
     level=DICT_GROUPBY_LEVELS[level],
 )
@@ -97,7 +103,7 @@
 cols = st.multiselect(
     f"Choose the {level.lower()} to display:",
     options=default_objs,
-    default=default_objs[1],
+    default=default_objs[0],
     key="sel_lev_1",
 )
 
@@ -118,7 +124,7 @@
 cols = col_l_lw.multiselect(
     f"Choose the {level.lower()} to display:",
     options=default_objs,
-    default=default_objs[1],
+    default=default_objs[0],
     key="sel_lev_2",
 )
 
@@ -132,6 +138,7 @@
 df_roll_ret = get_rolling_returns(
     df_prices=df_common_history.loc[first_day:last_day, :].ffill(),
     df_registry=df_registry,
+    tickers_to_evaluate=ticker_list,
     level=DICT_GROUPBY_LEVELS[level],
     window=window,
 )[cols]

src/pages/3_⚠️_Risk_Analysis.py

@@ -29,7 +29,12 @@
     st.error("Oops... there's nothing to display. Go through 🏠 first to load the data")
     st.stop()
 
-ticker_list = df_transactions["ticker_yf"].unique().tolist()
+df_n_shares = (
+    df_transactions.groupby("ticker_yf")
+    .agg(n_shares=("shares", "sum"))
+    .sort_values("n_shares", ascending=False)
+)
+ticker_list = df_n_shares.loc[~(df_n_shares == 0).all(axis=1)].index.unique().to_list()
 df_common_history = get_max_common_history(ticker_list=ticker_list)
 
 st.markdown("## Global settings")
@@ -73,6 +78,7 @@
 df_rets = get_period_returns(
     df=df_common_history.loc[first_day:last_day, :],
     df_registry=df_registry,
+    tickers_to_evaluate=ticker_list,
     period=DICT_FREQ_RESAMPLE[freq],
     level=DICT_GROUPBY_LEVELS[level],
 )
@@ -81,7 +87,7 @@
 cols = st.multiselect(
     f"Choose the {level.lower()} to display:",
     options=default_objs,
-    default=default_objs[1],
+    default=default_objs[0],
     key="sel_lev_3",
 )
 

src/plot.py

@@ -40,7 +40,7 @@ def plot_pnl_by_asset_class(
             hoverinfo="skip",
         )
     )
-    fig.update_traces(texttemplate="%{y:.1f}")
+    fig.update_traces(texttemplate="%{y:,.1f} €")
     fig.update_layout(
         autosize=False,
         height=550,

src/returns.py

@@ -11,9 +11,12 @@
 def get_period_returns(
     df: pd.DataFrame,
     df_registry: pd.DataFrame,
+    tickers_to_evaluate: list[str],
     period: Literal["Y", "Q", "M", "W", None],
     level: Literal["ticker", "asset_class", "macro_asset_class"],
 ):
+    # Filtra solo i ticker effettivamente in portafoglio
+    df_registry = df_registry[df_registry["ticker_yf"].isin(tickers_to_evaluate)]
     # Calcolo il ritorno
     df_rets = df.pct_change()[1:]
     # Se il periodo è None, la frequenza è giornaliera
@@ -29,7 +32,6 @@ def get_period_returns(
                 cols_to_sum = df_registry[df_registry[level].eq(class_)][
                     "ticker_yf"
                 ].to_list()
-
                 df_rets_classes[class_] = df_rets[cols_to_sum].sum(axis=1)
             return df_rets_classes
     # Se il periodo non è None, faccio resampling al periodo desiderato
@@ -52,9 +54,12 @@ def get_period_returns(
 def get_rolling_returns(
     df_prices: pd.DataFrame,
     df_registry: pd.DataFrame,
+    tickers_to_evaluate: list[str],
     window: int,
     level: Literal["ticker", "asset_class", "macro_asset_class"],
 ) -> pd.DataFrame:
+    # Filtra solo i ticker effettivamente in portafoglio
+    df_registry = df_registry[df_registry["ticker_yf"].isin(tickers_to_evaluate)]
 
     df_log_ret = np.log(df_prices.div(df_prices.shift(1)))
     df_roll_log_ret = df_log_ret.rolling(window=window).sum()

src/var.py

@@ -9,7 +9,7 @@
     os.path.join(*split_script_running_path[0 : len(split_script_running_path) - 1])
 )
 
-APP_VERSION = "0.2.0"
+APP_VERSION = "0.2.1"
 
 # Data/images