PDS/SBN New Horizons K4 peer review - T. Barnes noted issues =============================================== NOTE: (1) Third delivery of K4 L2/3 data sets. Last delivery as 2021-08-24. (2) These are the remaining issues from the internal review at SBN. =============================================== Datasets Validated: NH-A-ALICE-2-KEM1-V5.0 NH-A-ALICE-3-KEM1-V5.0 NH-A-LEISA-2-KEM1-V5.0 NH-A-LEISA-3-KEM1-V5.0 NH-A-LORRI-2-KEM1-V5.0 NH-A-LORRI-3-KEM1-V5.0 NH-A-MVIC-2-KEM1-V5.0 NH-A-MVIC-3-KEM1-V5.0 NH-A-PEPSSI-2-KEM1-V5.0 NH-A-PEPSSI-3-KEM1-V5.0 NH-A-REX-2-KEM1-V4.0 NH-A-REX-3-KEM1-V4.0 NH-A-SDC-2-KEM1-V5.0 NH-A-SDC-3-KEM1-V5.0 NH-A-SWAP-2-KEM1-V5.0 NH-A-SWAP-3-KEM1-V5.0 =============================================== Errors for: GLOBAL (all datasets) [NONE] =============================================== Errors for: ALICE [NONE] =============================================== Errors for: LEISA file: 'catalog/dataset.cat' --> In the version history section (perhaps in the "General statement about data set versions after V1.0" assuming this can happen anytime instead of just this particular version), discuss the changes to the data files that was reported in the data error below. The user should be made aware that the data files have been adjusted in XYZ manner due to ABC reason for the current version. directory: 'data/' --> Something odd about some of the LEISA data. I found a file (Ex: data/20181231_040860/lsb_0408605304_0x53c_eng.lbl) that included 2 new extensions since V4.0, "EXTENSION_WINDOW_MISMATCHES_*", yet the file description removes the text saying "This image may be windowed" and replaces it with "This image is not windowed". Why then the inclusion of these two new extensions? I found another instance (ex: nh-a-leisa-3-kem1-v5.0/data/20181231_040860/lsb_0408605304_0x53c_sci.lbl) where the text was changed, but no extensions were added, nor did they exist in V4.0. What is going on? I found similar situations scattered in both the raw and calibrated data sets. --> --> BLE 8/24: suggest NO CHANGE. Looks like this is all working well. The K3 level 2 FITS file was a 254x46x2 'window'. On DOY 2020/284, we played back the entire 256x256x285 file. The latest (full) file version is the one in K4. It is NO LONGER a windowed file. The new EXTENSION_WINDOW_MISMATCHES partition (and its header) get inserted when the pipeline detects some mismatches between the previous (windowed) and new (full) file data, due to compression or whatever. The mismatches partition in the new, non-windowed file allows a LEISA developer to go back and look into the differences if they so desire. It is probably of very little use to external users, but I recently suggested to the LORRI folks that these partitions should be removed from the PDS versions; no dice, they wanted them to remain. --> -> BLE 8/24: The reason we don't see these partitions in the LEISA level 3 files is that the mismatch tables are level-2-only, I.e. window mismatches are a 'raw file' phenomena. --> --> TFB 8/24: Okay, but I said I found similar situations in L3, not just L2. I only provided an example from L2. Here is an example in "nh-a-leisa-3-kem1-v5.0/data/20181231_040860/lsb_0408605304_0x53c_sci.lbl". The data parts appear to be unchanged, yet now the text says it is not windowed as before it said it "may be windowed" and proceeded to give the details of the windowing. =================== 119,134c119 < - This image may be windowed < - N.B. this may be a sliding window from DARK_SKY compression < and the following parameters may not describe the window < as it progresses through NAXIS3 (BANDS) of the cube. < Refer to the files in the DOCUMENT/ directory with names < like LEISA_WINDOW_*.LBL and LEISA_WINDOW_*.CSV, and to the < Operating Modes section of the LEISA instrument catalog < file CATALOG/LEISA.CAT. < - WINDOWX=46 (offset* to left edge of window) < - WINDOWY=254 (offset* to bottom edge of window) < - WINDOWW=34 (width of window) < - WINDOWH=2 (height of window) < - WIDTH =256 (width of image) < - HEIGHT =256 (height of image) < - * offsets are zero-based < - Pixels outside any windows will be set to arbitrary values --- > - This image is not windowed =================== --> --> TFB 8/24: In either event, these changes are not documented in the version history section in the dataset.cat file. The user should be made aware of such changes, as they are made aware of possible calibrated data changes due to geometry for pipeline re-runs for versions after v1.0. =============================================== Errors for: LORRI [NONE] =============================================== Errors for: MVIC files: '2021-08-24/nh-a-mvic-2-kem1-v5.0/calib/spectra/blue_new_chi.lbl' --> Recommend (but not required) that for each column NAME, that you not start the value on the next line. Currently column 2 still needs to be fixed. =============================================== Errors for: PEPSSI [NONE] =============================================== Errors for: REX [NONE] =============================================== Errors for: SDC [NONE] =============================================== Errors for: SWAP file: 'document/traj/traj*.tab/fmt' --> Three MISSING_CONSTANT value change from "-1000000.000" to "-0000000.000" for columns 7 thru 9. Why? It wasn't a problem before. I could imagine a negative zero might cause some problems on some systems, but it validates well.