Make timestamps consistent (#10)

* Initial changes

* Fix lint error

* Fix another lint error

Author: amd-pworfolk

src/lemonade/cache.py

@@ -1,5 +1,4 @@
 import os
-from datetime import datetime, timezone
 
 # Allow an environment variable to override the default
 # location for the build cache
@@ -17,20 +16,17 @@ def checkpoint_to_model_name(checkpoint_name: str) -> str:
     return checkpoint_name.split("/")[1]
 
 
-def get_timestamp() -> str:
+def get_build_timestamp(build_time) -> str:
     """
     Get a timestamp string in the format:
         <year>y_<month>m_<day>d_<hour>h_<minute>m_<second>s
     """
-    # Get the current time in GMT
-    current_time = datetime.now(timezone.utc)
-
     # Format the timestamp string
-    timestamp = current_time.strftime("%Yy_%mm_%dd_%Hh_%Mm_%Ss")
+    timestamp = build_time.strftime("%Yy_%mm_%dd_%Hh_%Mm_%Ss")
     return timestamp
 
 
-def build_name(input_name):
+def build_name(input_name, build_time):
     """
     Name the lemonade build by concatenating these two factors:
         1. Sanitize the input name (typically a model checkpoint name) by
@@ -54,7 +50,7 @@ def build_name(input_name):
         input_name_sanitized = input_name_sanitized.replace(":", "-")
 
     # Get the formatted timestamp string
-    timestamp = get_timestamp()
+    timestamp = get_build_timestamp(build_time)
 
     return f"{input_name_sanitized}_{timestamp}"
 

src/lemonade/cli.py

@@ -1,5 +1,6 @@
 import os
 import platform
+from datetime import datetime
 
 # pylint: disable=C0413
 # Prevent HF warnings from showing on every import
@@ -151,10 +152,12 @@ def main():
         first_tool_args.append("--input")
         first_tool_args.append(global_args["input"])
 
+        build_time = datetime.now().astimezone()
         state = State(
             cache_dir=os.path.abspath(global_args["cache_dir"]),
-            build_name=cache.build_name(global_args["input"]),
+            build_name=cache.build_name(global_args["input"], build_time),
             sequence_info=sequence.info,
+            build_time=build_time,
         )
         sequence.launch(state)
     else:

src/lemonade/common/filesystem.py

@@ -187,9 +187,13 @@ def get_available_builds(cache_dir):
 
     check_cache_dir(cache_dir)
 
+    builds_dir = os.path.abspath(build.builds_dir(cache_dir))
+    if not os.path.isdir(builds_dir):
+        return []
+
     builds = [
         pathlib.PurePath(build_name).name
-        for build_name in os.listdir(os.path.abspath(build.builds_dir(cache_dir)))
+        for build_name in os.listdir(builds_dir)
         if os.path.isdir(build.output_dir(cache_dir, build_name))
         and is_build_dir(cache_dir, build_name)
     ]
@@ -228,7 +232,7 @@ class Keys:
     # Prefix for reporting the execution status of a tool
     # In the report this will look like tool_status:TOOL_NAME
     TOOL_STATUS = "tool_status"
-    # Records the date and time of the evaluation after analysis but before
+    # Records the local date and time of the evaluation after analysis but before
     # build and benchmark
     TIMESTAMP = "timestamp"
     # Records the logfile of any failed tool/benchmark

src/lemonade/sequence.py

@@ -3,10 +3,8 @@
 import os
 import platform
 import copy
-from datetime import datetime
 from typing import List, Dict, Optional
 
-import pytz
 import psutil
 
 import lemonade.common.printing as printing
@@ -115,8 +113,7 @@ def launch(
         Executes the sequence of tools.
         """
 
-        current_time = datetime.now()
-        timestamp = current_time.strftime("%Y-%m-%d-%H%M%S")
+        timestamp = state.build_time.strftime("%Y-%m-%d-%H%M%S")
         start_times = {"warmup": time.time()}
 
         # Allow monitor to be globally disabled by an environment variable
@@ -161,11 +158,7 @@ def launch(
         state.save_stat(fs.Keys.BUILD_STATUS, build.FunctionStatus.INCOMPLETE)
 
         # Save a timestamp so that we know the order of builds within a cache
-        pacific_tz = pytz.timezone("America/Los_Angeles")
-        state.save_stat(
-            fs.Keys.TIMESTAMP,
-            datetime.now(pacific_tz),
-        )
+        state.save_stat(fs.Keys.TIMESTAMP, state.build_time)
 
         # Save the system information used for this build
         system_info = build.get_system_info_from_server()

src/lemonade/state.py

@@ -1,5 +1,6 @@
 import os
 import sys
+from datetime import datetime
 from typing import Dict, Optional, Any
 import yaml
 import lemonade.common.build as build
@@ -61,6 +62,7 @@ def __init__(
         cache_dir: str,
         build_name: Optional[str] = None,
         sequence_info: Dict[str, Dict] = None,
+        build_time: Optional[datetime] = None,
         **kwargs,
     ):
 
@@ -75,6 +77,9 @@ def __init__(
         self.cache_dir = parsed_cache_dir
         self.build_name = build_name
         self.sequence_info = sequence_info
+        self.build_time = (
+            datetime.now().astimezone() if build_time is None else build_time
+        )
         self.lemonade_version = lemonade_version
         self.build_status = build.FunctionStatus.NOT_STARTED
         self.downcast_applied = False

src/lemonade/tools/report/llm_report.py

@@ -5,6 +5,7 @@
 import re
 from typing import List
 import lemonade.common.printing as printing
+import lemonade.common.exceptions as exp
 import lemonade.common.filesystem as fs
 from lemonade.tools.management_tools import ManagementTool
 from lemonade.cache import DEFAULT_CACHE_DIR
@@ -148,7 +149,13 @@ def run(
     ):
         # Process input arguments
         cache_dirs = [os.path.expanduser(dir) for dir in input_caches]
-        cache_dirs = fs.expand_inputs(cache_dirs)
+        try:
+            cache_dirs = fs.expand_inputs(cache_dirs)
+        except exp.ArgError as e:
+            printing.log_info(
+                f"No Lemonade cache folders exist for: {cache_dirs}. Error: {e}"
+            )
+            return
         report_dir = os.path.expanduser(output_dir)
 
         if perf:

src/lemonade/tools/report/table.py

@@ -1,5 +1,5 @@
 from abc import ABC, abstractmethod
-from datetime import datetime, timezone
+from datetime import datetime
 import re
 from typing import Tuple, Dict, List
 import textwrap
@@ -585,9 +585,10 @@ def get_report_name(prefix: str = "") -> str:
         """
         Returns the name of the .csv report
         """
-        day = datetime.now().day
-        month = datetime.now().month
-        year = datetime.now().year
+        currentTime = datetime.now()
+        day = currentTime.day
+        month = currentTime.month
+        year = currentTime.year
         date_key = f"{year}-{str(month).zfill(2)}-{str(day).zfill(2)}"
         return f"{prefix}{date_key}.csv"
 
@@ -774,7 +775,7 @@ def include_stats(self, model_stats) -> bool:
         # Filter out build if it is too old
         if not self.days is None:
             build_day = model_stats[fs.Keys.TIMESTAMP]
-            today = datetime.now(timezone.utc)
+            today = datetime.now()
             delta = today - build_day
             if delta.days > self.days:
                 return False
@@ -891,7 +892,7 @@ def matching_builds(build_stats_1: Dict, build_stats_2: Dict) -> bool:
 
     @staticmethod
     def get_report_name() -> str:
-        current_time = datetime.now(timezone.utc)
+        current_time = datetime.now()
         timestamp = current_time.strftime("%Y-%m-%d-%H%M%S")
         return f"{timestamp}_perf.csv"