From 68798d807806a6d681af5b93d770fd4e23e9ce92 Mon Sep 17 00:00:00 2001
From: Nicholas Shavers <nshavers@contractor.usgs.gov>
Date: Thu, 5 Dec 2024 15:51:48 -0800
Subject: [PATCH] with compression at the file level
---
geomagio/ImagCDFFactory.py | 19 +++++++++++++------
1 file changed, 13 insertions(+), 6 deletions(-)
diff --git a/geomagio/ImagCDFFactory.py b/geomagio/ImagCDFFactory.py
index 157ad0cf..0acf6ece 100644
--- a/geomagio/ImagCDFFactory.py
+++ b/geomagio/ImagCDFFactory.py
@@ -129,12 +129,20 @@ class ImagCDFFactory(TimeseriesFactory):
def write_file(self, fh, timeseries: Stream, channels: List[str]):
# Create a temporary file to write the CDF data
- with tempfile.NamedTemporaryFile(delete=False) as tmp_file:
- tmp_file_path = tmp_file.name + ".cdf"
+ with tempfile.NamedTemporaryFile(delete=False, suffix='.cdf') as tmp_file:
+ tmp_file_path = tmp_file.name
try:
# Initialize the CDF writer
- cdf_writer = cdflib.cdfwrite.CDF(tmp_file_path)
+ cdf_spec = {
+ 'Compressed': 9, # Enable compression (0-9)
+ # 'Majority': 'row_major', # Data layout - gets set automatically
+ # 'Encoding': 'ibmpc', # Corrupts CDF - gets set automatically
+ 'Checksum': True, # Disable checksum for faster writes (optional)
+ 'rDim_sizes': [], # Applicable only if using rVariables - CDF protocol recommends only using zVariables.
+ }
+
+ cdf_writer = cdflib.cdfwrite.CDF(path=tmp_file_path, cdf_spec=cdf_spec, delete=True)
# Write global attributes
global_attrs = self._create_global_attributes(timeseries, channels)
@@ -206,7 +214,6 @@ class ImagCDFFactory(TimeseriesFactory):
# Cleanup the temporary file
os.remove(tmp_file_path)
-
def put_timeseries(
self,
timeseries: Stream,
@@ -288,7 +295,7 @@ class ImagCDFFactory(TimeseriesFactory):
# Read existing data to merge with new data
existing_cdf = cdflib.cdfread.CDF(url_file)
existing_stream = self._read_cdf(existing_cdf)
- existing_cdf.close()
+ # existing_cdf.close() #no close method?
existing_data = existing_stream
# Merge existing data with new data
@@ -480,7 +487,7 @@ class ImagCDFFactory(TimeseriesFactory):
elif channel in TEMPERATURE_ELEMENTS_ID:
units = 'Celsius'
fieldnam = f"Temperature {temperature_index} {trace.stats.location}"
- validmin = -273.15
+ validmin = -273.15 #absolute zero
validmax = 79_999
elif channel in ['F','S']:
units = 'nT'
--
GitLab