Changes

Yun-Jing Huang · 49ebda9b
--- a/240206.md
+++ b/240206.md
@@ -70,3 +70,356 @@ It would also be very helpful to us to know what  (in detail) the whole gstlal p


 ## Minutes
+## Agenda / Minutes
+
+* Announcements (5 minutes)
+  - Please check the rota for next week's call
+  - Confirmation of next week's focus session
+    - Cort will take over after Jim leaves
+* Urgent Dev Discussion (5 - 10 min)
+  - Merging banks together [MR](https://git.ligo.org/chad-hanna/manifold/-/merge_requests/36)
+    - Jim: Urgent for getting SSM to production
+      - combine banks, ready to ship
+    - Victoria: Shio and I want to put pastro for online ssm for o4b, will confirm with LL team if pastro is needed, if so, will work with Anarya to get that ready.
+
+* Focus Session:
+  - Answer Roberto's Questions (see below section)
+* [Last week's East call](https://wiki.ligo.org/CBC/Searches/GstLALEastAgenda20240131)
+* Quick updates (45 minutes)
+  - Operations (5 minutes)
+    - LL CBC operations
+      - Shomik: Prat and I are working on getting stuff ready for Esme.
+  - O4 Dev (30 minutes)
+    - Low latency integrated testing and Monitoring
+      - Shomik: Skymap test suite jobs failed a few times
+      - Cort: Removing getenv for test suite
+    - O4b online production lead team updates ([issue](https://git.ligo.org/gstlal/projects/-/issues/795 "Assemble the O4b MDC team"))
+      - Engineering run plans
+        - Shomik: working on container and Makefile
+      - AllSky
+        - Shomik: have some event plotter errors, might be snr optimizer error with dtdphi plots
+      - EW
+        - Amanda: EW is fine, haven't seen errors. Should we remove counts for EW for all events?
+          - Shomik: should for consistency
+          - Amanda: but in o4a, EW is not collecting background
+          - Leo: should remove counts, will remove some leaked in signal, or it will do nothing
+      - SSM
+      - IMBH
+        - Debnandini: Prathamesh did likelihood and count studies. Found extinction model is not able to fix the bump, will run on a new mass model.
+        - Anarya: is it using a manifold bank?
+        - Debnandini: Prathamesh: seed likelihood mass model
+        - Anarya: explains seed likelihood mass model, doesn't have to do with manifold, do you have plots for the mass model?
+          - Debnandini: not for mass model
+        - Anarya: improve mass model in source frame instead of detector frame, to account for redshift, could try that
+        - Debnandini: the seed thing is in the rerank step?
+          - Anarya: yes
+        - Debnandini: is that a separate h5 file?
+        - Anarya: yes
+        - Debnandini: take discussion offline, we are working on the mass model and fixing the run
+    - GWTC-4
+    - Computing
+      - Cort: on ICDS, have default value for X509_USER_PROXY, but not on other clusters, suggest removing it
+    - Likelihood ratio, background and foreground sampling
+    - Injection file format
+    - SNR optimizer
+    - offline DAG and Ops
+      - Amanda: offline injection filtering is almost done, moved to icds, half of Edward is on CIT and half of Edward is run on icds
+      - Prat: try to run on all the directories, ranking stats can account for all the files
+        - Leo: what to you mean by all the directories?
+        - Prat: put all the chunks together
+        - Cort: you can combine result of the filter dags
+        - Prat: concerned with Virgo
+        - Leo: ranking code, it's great that it is working, is there a subset injection set run?
+        - Leo: it's ok if there is only one injection set, Anarya has offline pastro workflow, we need for GWTC, if you have injection recovery results for a subset of chunks, pass it to Anarya.
+        - Prat: will have results soon, will pass to Anarya
+      - Victoria: offline SSM: settled on the offline bank for SSM, will work with Anarya to make the mass model for the offline bank, to run jobs
+        - Cort: will help in determining which cluster to run on
+
+  - Exploratory development/Misc projects (5 minutes)
+- Paper Updates
+  - Assigning sprint tasks to those who were not at sprint.
+* AOB
+
+
+## Roberto's Questions
+
+It would also be very helpful to us to know what  (in detail) the whole gstlal plan for O4b
+
+  1. in which computer center are running the search and how many instances of them
+  2. the expected number of trigger upload for each real high-SNR events.
+  3. How will VIRGO data be used for assessing SIGNIFICANCE/FAR and skymaps?
+  4. Do you plan to upload single detector VIRGO trigger?
+  5. Will the single detector trigger be uploaded for SSM, allsky, and EarlyWarning triggers?
+  6. What is the single detector FAR threshold for single detector uploads ?
+  7. Whould you confirm that the triggers will be marked HLV when PSD information for VIRGO (and indeed, I will be possible to generate HLV skymaps)  are presents?
+  8. How can we derive the information of which detector contributed to the significance of a trigger. Example of: [https://gracedb.ligo.org/superevents/S231020bw/view/](https://gracedb.ligo.org/superevents/S231020bw/view/ "https://gracedb.ligo.org/superevents/S231020bw/   --->view/") so that [https://gracedb.ligo.org/events/G447745/view/](https://gracedb.ligo.org/events/G447745/view/ "https://gracedb.ligo.org/events/G447745/view/") correspond to an event whose significance was derived by the signal in L1 ?  H1 report χ2=None.
+
+  - Rachael: response in google doc: https://docs.google.com/document/d/18Dzuqhgc_lyV31TnUsPpQVXn6B8CW_43jZUBM7J36-Q/edit?usp=sharing
+    - Have been discussing with Roberto about injection streamer, they want our full o4b plan
+    - 2. for all sky, 6 uploads?
+  - Shio: Shomik has already answered the questions: link: https://chat.ligo.org/ligo/pl/bkfrad67h7bx9pxdrhc14i1h1a
+  - Leo: Real high-SNR events?
+    - Rachael: non inj events
+  - Rachael: Want to talk about virgo?
+    - Zach: not been able to get much progress, waiting for Prathamesh to recover
+  - Rachael: is EW uploading single detectors in o4b?
+    - Surabhi: No
+    SSM?
+    - Victoria: idk
+    - Shomik: we should, same as o4a
+  - Rachael: FAR threshold for single detector?
+    - Shomik: same as other uploads
+  - Rachael: 7: normal filtering Virgo, marks HLV, if using V for snr optimizer, HL?
+    - Shomik: if snr optimizer, initially HL, but if there are V for snr optimizer, there will be HLV
+    - Leo: snr optimizer, will always eventually have HLV?
+    - Shomik: snr optimizer will always come before the second alert goes out, will mostly come within 40 sec, only one past 5 min. https://ldas-jobs.gwave.ics.psu.edu/grafana/d/tFWp595Sz/shomik-test-suite-with-optimizer?orgId=1&refresh=1m&var-DashDatasource=performance_o4b&var-far_threshold=1.23e-4&var-pastro_name=gstlal&var-far=0.0001230000000000&var-Display_Optimizer_Results=%22snr_optimized%22%3D%27False%27%20OR%20%22snr_optimized%22%3D%27None%27%20OR%20%22snr_optimized%22%3D%27True%27&from=1704220782000&to=now&viewPanel=14
+    - Leo: should make it clear the distinction between first and second alert
+  - Surabhi: 8. Determining whether Virgo can help with skymaps
+    - Leo: if contribute to network snr, will contribute to p(ifo). If snr < 4, neglibile contribution, if ifo is off, no contribution. Depends on whether we filter Virgo or not. Even if optimizer adds Virgo, it does not contribute to the significance.
+    - Rachael: So if use optimizer for Virgo, it won't recalculate significance, right?
+      - Zach: right, currently not planning for optimizer to do anything with significance.
+      - Cody: Does optimizer compute a chisq?
+        - Shomik: yes
+      - Cody: write in user guide, we don't use Virgo for significance
+      - Shomik: every other pipeline have decided that they are not using Virgo significance, only skymaps
+      - Becca and Rachael: I thought we said last week, other pipelines are going to filter Vrigo
+      - Shomik: That is what Chad wants. Other pipelines will use Virgo for SNR(?). We will not be preferred, because we only have HL not HLV for the first alert.
+      - Cody: I thought LL group will ignore Virgo, so Chad's concern is not warranted
+      - Rachael: if what Cody said is true, what do we prefer?
+        - Shomik and Zach: prefer Virgo in optimizer
+      - Cody: if Virgo pass cut off, prefer to filter HLV
+      - Becca: What the reason for not filtering Virgo
+        - Shomik: worried about gap between sensitivity between HL and V
+      - Surabhi: Do we still have the HV LV single commit?
+        - Shomik: that commit is reverted
+        - Surabhi: But we will not have singles penalty
+      - Shomik: will work on looking at MDC11 to check for number of singles
+      - Cody: since sensitivity difference is similar to O3, we should do that in O3
+      - Leo: Since we have count tracker, it can fix the issue
+      - Cody: someone needs to write down at what date we need a fresh run to test this.
+      - Rachael: concerned that if we haven't run on MDC with a new code version, it will not meet the minimum requirement, I think we need to run on at least a full MDC.
+      - Shomik: Roberto mentioned only need to run on 1 week of MDC. Only need to pass SCCB
+      - Rachael: Thought Patrick Sutton mentioned we need to run on MDC
+      - Leo: Should study first if count tracker works, then will need to have unrealistic review schedule.
+      - Rachael: See decision tree from o4b team
+      - Shomik: Can come up with the information
+      - Rachael: revisit on Thur scrum, will hold off from responding to Roberto
+
+
+
+
+## Chat log
+Shio Sakon to Everyone (Feb 6, 2024, 1:35 PM)
+Oh did I?
+ 
+Cort Posnansky to Everyone (Feb 6, 2024, 1:35 PM)
+I can do it if needed
+ 
+Shio Sakon to Everyone (Feb 6, 2024, 1:36 PM)
+I haven’t been able. To look at it yet. Will do this afternoon
+ 
+Rachael Huxford to Everyone (Feb 6, 2024, 1:39 PM)
+https://docs.google.com/document/d/18Dzuqhgc_lyV31TnUsPpQVXn6B8CW_43jZUBM7J36-Q/edit?usp=sharing
+ 
+Shio Sakon to Everyone (Feb 6, 2024, 1:40 PM)
+I thought shomik already answered in mattermost?
+ 
+Shio Sakon (Feb 6, 2024, 1:40 PM)
+https://chat.ligo.org/ligo/pl/bkfrad67h7bx9pxdrhc14i1h1a
+ 
+jim to Everyone (Feb 6, 2024, 1:41 PM)
+sure
+ 
+Shio Sakon to Everyone (Feb 6, 2024, 1:44 PM)
+same
+
+jim to Everyone (Feb 6, 2024, 1:44 PM)
+Classic gdoc perms
+
+Patrick Godwin to Everyone (Feb 6, 2024, 1:45 PM)
+can you zoom in a bit?
+ 
+Shio Sakon (Feb 6, 2024, 1:45 PM)
+https://docs.google.com/document/d/18Dzuqhgc_lyV31TnUsPpQVXn6B8CW_43jZUBM7J36-Q/edit?usp=sharing
+@jim can you zoom in?
+ 
+jim (Feb 6, 2024, 1:46 PM)
+better?
+ 
+Amanda Christine Baylor (Feb 6, 2024, 1:46 PM)
+More please
+perfect
+ 
+Patrick Godwin (Feb 6, 2024, 1:47 PM)
+looks good
+ 
+jim (Feb 6, 2024, 1:47 PM)
+Can turn up to 11
+ 
+Shio Sakon to Everyone (Feb 6, 2024, 1:51 PM)
+Like within 5 min, right?
+ 
+Shio Sakon (Feb 6, 2024, 1:51 PM)
+Snr opt
+
+Shomik Adhicary to Everyone (Feb 6, 2024, 1:51 PM)
+https://ldas-jobs.gwave.ics.psu.edu/grafana/d/tFWp595Sz/shomik-test-suite-with-optimizer?orgId=1&refresh=1m&var-DashDatasource=performance_o4b&var-far_threshold=1.23e-4&var-pastro_name=gstlal&var-far=0.0001230000000000&var-Display_Optimizer_Results=%22snr_optimized%22%3D%27False%27%20OR%20%22snr_optimized%22%3D%27None%27%20OR%20%22snr_optimized%22%3D%27True%27&from=1704220782000&to=now&viewPanel=14
+
+Cody Messick (Feb 6, 2024, 2:00 PM)
+I wasn't here for any of the discussion, so all I have is this:
+
+https://git.ligo.org/emfollow/gwcelery/-/issues/748
+
+and the fact that the gwcelery librarians are working on implementing a manual button to generate HLV sky maps
+ 
+Cody Messick to Everyone (Feb 6, 2024, 1:57 PM)
+so even if the optimized SNR uploads before the second preliminary goes out
+only HL will be used for sky maps until someone manually pushes a button
+ 
+jim to Everyone (Feb 6, 2024, 1:57 PM)
+I need to head out now, @Cort Posnansky can you screen share?
+ 
+Cort Posnansky (Feb 6, 2024, 1:57 PM)
+Yeah
+ 
+Zach Yarbrough to Everyone (Feb 6, 2024, 1:59 PM)
+It would have the same significance as the HL event
+ 
+Shio Sakon to Everyone (Feb 6, 2024, 2:01 PM)
+FAR and LR of the snr opt event is the same as the original right?
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:02 PM)
+Did I miss something though? I thought last week we said V1 was sensitive enough that we would filter and include in LR
+
+Cody Messick to Everyone (Feb 6, 2024, 2:03 PM)
+that information has not made its way to the LL group
+so someone should bring it up at the LL call tomorrow
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:03 PM)
+Yeah cody, I just meant we discussed last week on the dev call
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:04 PM)
+But I agree we just need to decide and bring it to the wider group
+ 
+Zach Yarbrough to Everyone (Feb 6, 2024, 2:04 PM)
+Optimizer does upload a V1 chisq btw
+ 
+Cody Messick to Everyone (Feb 6, 2024, 2:06 PM)
+im on eduroam and it sucks ass so I lost most of what zach said
+ 
+Shomik Adhicary (Feb 6, 2024, 2:07 PM)
+Check whether V is being treated as K or HL in the preferred event rule
+
+Becca Ewing to Everyone (Feb 6, 2024, 2:08 PM)
+But we originally said it should depend on sensitivity right?
+ 
+Zach Yarbrough to Everyone (Feb 6, 2024, 2:08 PM)
+Filter HLV
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:08 PM)
+We had a cut off
+ 
+Shio Sakon (Feb 6, 2024, 2:08 PM)
+I think we said 50 mpc
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:08 PM)
+Like 50 Mpc or something
+ 
+Patrick Godwin to Everyone (Feb 6, 2024, 2:09 PM)
+filter HLV, we can do it
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:09 PM)
+I say we hear what they say on ops this week and then just go with it
+What is the reason for *not* filtering HLV
+Sorry can’t open mic now
+So no lol
+ 
+Cody Messick to Everyone (Feb 6, 2024, 2:10 PM)
+so worried about like background collection with coincs
+like the issues in O3?
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:11 PM)
+Yeah I would say use that patch + filter HLV
+
+Shomik Adhicary (Feb 6, 2024, 2:13 PM)
+Got this issue to track all concerns. Please put comment, concerns, etc in there whenever you get an idea
+ 
+Leo-iPad to Everyone (Feb 6, 2024, 2:14 PM)
+agreed
+ 
+Zach Yarbrough to Everyone (Feb 6, 2024, 2:15 PM)
+Yeah that’s fair
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:15 PM)
+If ER is in 3 weeks I would say that cutoff is like 1-2 weeks from now
+ 
+Rachael Huxford to Everyone (Feb 6, 2024, 2:15 PM)
+The ER isn’t until midof march afaik
+So its like 5 weeks
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:16 PM)
+Actually that’s a good point Rachael
+ 
+Cody Messick to Everyone (Feb 6, 2024, 2:16 PM)
+yeah, but sccb requires running, doesn't it?
+ 
+Zach Yarbrough to Everyone (Feb 6, 2024, 2:17 PM)
+Patrick
+ 
+Cody Messick to Everyone (Feb 6, 2024, 2:17 PM)
+sutton
+ 
+Becca Ewing to Everyone (Feb 6, 2024, 2:17 PM)
+Safe bet for names is always patrick
+ 
+Surabhi Sachdev to Everyone (Feb 6, 2024, 2:17 PM)
+I like “Man with the nice moustache”
+ 
+Surabhi Sachdev to Everyone (Feb 6, 2024, 2:18 PM)
+sounds like a mystery novel
+ 
+Cody Messick to Everyone (Feb 6, 2024, 2:18 PM)
+okay i was wrong, 36 days until engineering run
+ 
+Surabhi Sachdev (Feb 6, 2024, 2:18 PM)
+we don’t know yet for sure though 36+-5?
+or more like +5/6?
+ 
+Cody Messick (Feb 6, 2024, 2:19 PM)
+ah gotcha
+ 
+Cody Messick to Everyone (Feb 6, 2024, 2:18 PM)
+March 13, at least that was the originally announced date
+ 
+Shomik Adhicary (Feb 6, 2024, 2:19 PM)
+Yup, they gonna push that back by <1 week
+
+Cody Messick to Everyone (Feb 6, 2024, 2:21 PM)
+i havent felt good about something since primary school
+ 
+Surabhi Sachdev (Feb 6, 2024, 2:22 PM)
+me too Cody
+
+Shomik Adhicary to Everyone (Feb 6, 2024, 2:26 PM)
+I would recommend removing counts for consistency with O4a
+ 
+Shomik Adhicary (Feb 6, 2024, 2:27 PM)
+Preneting signal contamination is a good idea
+ 
+Victoria Niu (Feb 6, 2024, 2:27 PM)
+Can we design a robot to do it for us?
+ 
+Shomik Adhicary (Feb 6, 2024, 2:29 PM)
+Glad you volunteered yourself to automate this 🙃
+ 
+Victoria Niu (Feb 6, 2024, 2:30 PM)
+What can I say. I am cheaper than robot
+ 
+Shio Sakon to Everyone (Feb 6, 2024, 2:29 PM)
+Is m2 = 0.8 Msun for imbh mass model?
+
+Cody Messick to Everyone (Feb 6, 2024, 2:33 PM)
+I gotta run, I might be back, i might not. Ttyl
+
+Amanda Christine Baylor to Everyone (Feb 6, 2024, 2:37 PM)
+Different injection sets
\ No newline at end of file