Pipeline-Testing.md

@@ -2,8 +2,7 @@ Pipeline tests page.
 
 # Testing on Simulated data
 
-- Recovery of injection in no noise dataset:
-No noise dataset from the pyGWB simulator module. Injection of power law ⅔ with Omega=1.09e-8 at fref = 25 Hz. Simulated data has a sampling frequency of 1024 Hz, flow and fhigh have to be adjusted accordingly compared to actual data.
+- Recovery of injection in no noise dataset: No noise dataset from the pyGWB simulator module. Injection of power law ⅔ with Omega=1.09e-8 at fref = 25 Hz. Simulated data has a sampling frequency of 1024 Hz, flow and fhigh have to be adjusted accordingly compared to actual data.
 
 Ran a condor with job length of 10000s, data can be found in /home/max.lalleman/public_html/Code/GWB_Try/pygwb/DAG/output/Simulated_Data_Highpass
 
@@ -11,24 +10,22 @@ You can find the notebook that calculates the overall point estimate at /home/ma
 
 The injected Omega was recovered, pyGWB gave a value of 1.08933e-8.
 
-- One Week dataset, SNR = 5:
-We ran both stochastic.m and pyGWB on a one week dataset. 
-See slide https://docs.google.com/presentation/d/1zI925LxacaHkhLeOdSMpPiQYO2nq1HlbDJ5tawu1gQ0/edit?pli=1#slide=id.g11786a4c73c_0_25 for comparison with stochastic.m (and a test for matching vs non-matching job duration for the simulator module).
+- One Week dataset, SNR = 5: We ran both stochastic.m and pyGWB on a one week dataset. See slide https://docs.google.com/presentation/d/1zI925LxacaHkhLeOdSMpPiQYO2nq1HlbDJ5tawu1gQ0/edit?pli=1#slide=id.g11786a4c73c_0_25 for comparison with stochastic.m (and a test for matching vs non-matching job duration for the simulator module).
 
 Data can be found in /home/max.lalleman/public_html/Code/GWB_Try/pygwb/DAG/output/Simulated_Data_One_Week_SNR_5, notebook is /home/max.lalleman/public_html/Code/GWB_Try/pygwb/DAG/Simulated_Data_One_Week_SNR5.ipynb
 
--Half Year dataset, SNR = 5
+\-Half Year dataset, SNR = 5
 
 - 100 day dataset
 
-
 # Testing on O3 data
 
-- 900 seconds: Our good old notebook shows that for small datasets we have recover at the 0.2% level both the point estimate and sigma compared to stochastic.m 
-
-- One day: Tested one day of O3 data and compared with stochastic.m on that dataset. Both notching and delta-sigma cut were implemented.  
-Sigmas agreed very well, while point estimates were a bit off, but due O3 data being noise and point estimates being in one sigma of each other, this is of little concern. 
-Data can be found in /home/max.lalleman/public_html/Code/GWB_Try/pygwb/DAG/output/Simulated_Data_100_day_segment. Notebook that analyses the data at /home/max.lalleman/public_html/Code/GWB_Try/pygwb/DAG/Simulated_Data_100_day.ipynb
+- 900 seconds: This [review notebook](https://git.ligo.org/pygwb/pygwb_review/-/blob/main/pipeline/pipeline_test_O3_data.ipynb) shows that for small datasets we have recover at the 0.2% level both the point estimate and sigma compared to stochastic.m
+- One day: Tested one day of O3 data and compared with stochastic.m on that dataset. Both notching and delta-sigma cut were implemented.\
+  Sigmas agreed very well, while point estimates differ, but due O3 data being noise and point estimates being in one sigma of each other, this is of little concern. Data can be found in /home/max.lalleman/public_html/Code/GWB_Try/pygwb/DAG/output/Simulated_Data_100_day_segment. Notebook that analyses the data at /home/max.lalleman/public_html/Code/GWB_Try/pygwb/DAG/Simulated_Data_100_day.ipynb
+- Full O3A+O3B analysis on <span dir="">DCS-CALIB_STRAIN_CLEAN_C01</span> data
+- Full O3A+O3B analysis on <span dir="">DCS-GATED_STRAIN_C01</span> data
+- Full O3A+O3B analysis on<span dir="">  DCS-CALIB_STRAIN_GATED_SUB60HZ_C01</span> data
 
 # Speed of pyGWB compared to Stochastic.m
 
@@ -40,4 +37,3 @@ Data can be found in /home/max.lalleman/public_html/Code/GWB_Try/pygwb/DAG/outpu
 - Running on half-year datasets from MDC (and pyGWB simulator module)
 - Investigating the speed of pyGWB for real O3 data.
 - What is the best length for job files?
\ No newline at end of file
-