Refactor RCCL log parser to enhance transfer reporting

- Rename transfer fields for clarity and introduce separate methods for reporting non-P2P and P2P transfers.
- Add new P2P fields extraction and sorting logic to improve data presentation.
- Update method names and comments for better understanding of functionality.

projects/rccl-log-parser/rccl_log_parser.py

@@ -23,7 +23,7 @@ class RcclLogParser:
             r"Disabled GDRCopy": "GDRCopy: disabled",
         }
 
-        # Pattern -> replacement
+        # Pattern -> column
         self.graph_info_fields = {
             r"Pattern": "Pattern",
             r"crossNic": "crossNic",
@@ -33,8 +33,8 @@ class RcclLogParser:
             r"sameChannels": "sameChannels",
         }
 
-        # Pattern -> replacement
-        self.transfer_fields = {
+        # Pattern -> column
+        self.cl_transfer_fields = {
             r"protocol": "protocol",
             r"nbytes": "nbytes",
             r"algorithm": "algorithm",
@@ -45,6 +45,17 @@ class RcclLogParser:
             r"chunksize": "chunksize",
         }
 
+        # Pattern -> column
+        self.p2p_fields = {
+            r"p2p : rank": "local",
+            r"send rank": "send",
+            r"recv rank": "recv",
+            r"p2pnChannelsPerPeer": "p2pnChannelsPerPeer",
+            r"p2pnChannels": "p2pnChannels",
+            r"nChannelsMax": "nChannelsMax",
+            r"protocol": "protocol",
+        }
+
     def collect(self, line):
         self.raw_lines.add(line)
 
@@ -58,7 +69,8 @@ class RcclLogParser:
         self._report_sys()
         self._report_user_envs()
         self._report_graph_info()
-        self._report_transfers()
+        self._report_cl_transfers()
+        self._report_p2p_transfers()
 
         print(" End of Report ".center(80, "="))
 
@@ -127,9 +139,9 @@ class RcclLogParser:
         print(extracted_df.fillna("-").to_string(index=False))
         print()
 
-    def _report_transfers(self):
-        """Extract transfer arguments"""
-        print("===> Unique Transfers:\n")
+    def _report_cl_transfers(self):
+        """Extract non-P2P transfer arguments"""
+        print("===> Unique Ring/Tree Transfers:\n")
 
         # Filter lines by looking for 'protocol' and 'nbytes'
         raw_lines = [
@@ -143,7 +155,7 @@ class RcclLogParser:
         df = pd.DataFrame(raw_lines, columns=["raw_log"])
 
         # Extract all fields using a single loop
-        for pattern, col_name in self.transfer_fields.items():
+        for pattern, col_name in self.cl_transfer_fields.items():
             df[col_name] = df["raw_log"].str.extract(
                 rf"\b{pattern}\s+(\S+)", expand=False
             )
@@ -161,12 +173,65 @@ class RcclLogParser:
         df.drop(columns=["raw_log"], inplace=True)
         df.drop_duplicates(inplace=True)
 
-        desired_order = ["nbytes", "protocol", "nchannels"]
-        sort_cols = [col for col in desired_order if col in df.columns]
+        sort_cols = ["nbytes", "protocol", "nchannels"]
+        sort_cols = [col for col in sort_cols if col in df.columns]
+
+        if sort_cols:
+            df.sort_values(by=sort_cols, inplace=True)
+
+        # Fill NaNs with "-" and print
+        print(df.fillna("-").to_string(index=False))
+        print()
+
+    def _report_p2p_transfers(self):
+        """Extract P2P transfer details"""
+        print("===> Unique P2P Transfers:\n")
+
+        # Filter lines by looking for 'p2p :' and 'send rank'
+        raw_lines = [
+            line for line in self.output if "p2p :" in line and "send rank" in line
+        ]
+
+        if not raw_lines:
+            print("  (No P2P transfers found)\n")
+            return
+
+        # Extract all fields using a single loop
+        df = pd.DataFrame(raw_lines, columns=["raw_log"])
+        for pattern, col_name in self.p2p_fields.items():
+            df[col_name] = df["raw_log"].str.extract(
+                rf"{pattern}\s+(\S+)", expand=False
+            )
 
+        # Type conversion for correct sorting
+        numeric_cols = [
+            "local",
+            "send",
+            "recv",
+            "p2pnChannelsPerPeer",
+            "p2pnChannels",
+            "nChannelsMax",
+        ]
+        for col in numeric_cols:
+            if col in df.columns:
+                df[col] = pd.to_numeric(df[col], errors="coerce")
+
+        # Clean up
+        df.drop(columns=["raw_log"], inplace=True)
+        df.drop_duplicates(inplace=True)
+
+        sort_cols = ["protocol", "local", "send", "recv"]
+        sort_cols = [c for c in sort_cols if c in df.columns]
         if sort_cols:
             df.sort_values(by=sort_cols, inplace=True)
 
+        # Move 'protocol' to the first column
+        cols = df.columns.tolist()
+        if "protocol" in cols:
+            cols.remove("protocol")
+            cols.insert(0, "protocol")
+            df = df[cols]
+
         # Fill NaNs with "-" and print
         print(df.fillna("-").to_string(index=False))
         print()