me

quietlight
May 23, 2026, 1:07 AM
YG3HYFC7QCOYC2HURVZNS64JBI3UUBNEJD5FE6JHJVIQGHKWHUSQC

Dependencies

  • [2] AWE634BL me
  • [3] QVIGQOQZ more work on utils/ with glm
  • [4] T6CPEI47 consolidated shared types from calls_from_preds.go into calls_from_common.go
  • [5] AISACMH6 me
  • [6] 2Y5U3QPU added gosymdb
  • [7] 43TMU2JO more tests, glm much better than claude
  • [8] LQLC7S3A trying gemini: Inconsistent Standards in @utils/ refactoring
  • [9] XO5DF6WR I tiedied up readme
  • [10] KZKLAINJ run out of space on nest, cleaned out
  • [11] P6HBSK3W me
  • [12] VNFPBXF7 moved dep tests to golangci-lint
  • [13] R4VYSIUK me
  • [14] 54GPBNIX added +_ for tui to select segments with no calltype

Change contents

  • replacement in me.txt at line 3
    [3.726623][3.118:133](),[3.133][3.2079:2107](),[3.726623][3.2079:2107]()
    tomtit cert 70
    +_ species (ie no calltype)
    [3.726623]
    [3.121]
  • replacement in me.txt at line 5
    [3.134][3.134:181](),[3.181][3.726638:726639](),[3.726638][3.726638:726639]()
    silvereye/pipipi
    rifleman/tomtit+contact/Tieke
    [3.134]
    [3.726639]
    rifleman/tomtit+contact/Tieke/identify
  • edit in me.txt at line 9
    [3.183]
    [3.726751]
    Dont Know should be cert 0, not 70: Julia code is not correct, But I maky have fixed it as some is fine
  • edit in me.txt at line 15
    [3.726787]
    [3.726787]
    os.Exit
  • edit in me.txt at line 24
    [3.726922][3.726922:727051]()
    to one hot encoded csv for opensoundscape (because python is so slow, and I would have to convert to raven selection.txt first)
  • edit in me.txt at line 43
    [3.727602][3.727602:727603](),[3.727603][3.2108:2153]()
    npm install -g @mariozechner/pi-coding-agent
  • edit in me.txt at line 50
    [3.350][3.727874:728003](),[3.727874][3.727874:728003]()
    pi-specific are in ~ somewhere (ok because keeps them seperate) if installed with eg: $pi install npm:@tmustier/pi-ralph-wiggum
  • replacement in me.txt at line 211
    [3.733898][2.253:312]()
    jsonfile = "$item/segment_summary_2026-05-15.json"
    [3.733898]
    [3.733957]
    jsonfile = "$item/segment_summary_2026-05-20.json"
  • edit in me.txt at line 246
    [3.734826][3.734826:734863]()
    ollama run qwen3.5:9b # uninstalled
  • replacement in me.txt at line 251
    [3.734921][3.734921:734960]()
    R620/2024-05-06 only
    Run Through Gemma
    [3.734921]
    [3.734960]
    (base) 2024-01-03 ➤ ollama list
    NAME ID SIZE MODIFIED
    qwen3.5:9b 6488c96fa5fa 6.6 GB 8 days ago
    ministral-3:14b 4760c35aeb9d 9.1 GB 8 days ago
    gemma4:26b 5571076f3d70 17 GB 8 days ago
    granite4.1:8b 444af1c4b2fe 5.3 GB 9 days ago
    gemma4:e4b c6eb396dbd59 9.6 GB 4 weeks ago
  • edit in me.txt at line 363
    [3.742717][3.742717:742718]()
  • edit in me.txt at line 365
    [3.742720][3.742720:742865]()
    Lets manually execute this loop once, when we are happy we will design a ralph loop together to loop through remaining BirdNET classes /grill-me
  • edit in me.txt at line 366
    [3.742866][3.742866:743209]()
    # Retrieve BirdNET List from folder /media/david/SSD4/Twenty_Four_Seven/R620/2024-05-06/
    skraak calls summarise --folder /media/david/SSD4/Twenty_Four_Seven/R620/2024-05-06/ --brief --filter BirdNET 2>/dev/null | jq -r '.filters.BirdNET.species | to_entries | map(select(.key | test("^[A-Z]"))) | sort_by(.value) | .[] | "\(.value)\t\(.key)"'
  • edit in me.txt at line 367
    [3.743210][3.743210:744105]()
    Start from the top of the BirdNET List and attempt to label a BirdNET class with one of the classes below using skill /call-classification, /call-classification-ollama, /call-library
    While there is only a few segments in the BirdNET class, attempt to do this yourself reading data from /call-classification, /call-library. It is your role to work out what this class actually is (BirdNET mislabels many New Zealand Birds), and to assign correct labels wherever possible
    When there are many segments in a BirdNET class use skill /call-classification-ollama. choose your reference images carefully. It is your role to work out what this class actually is (BirdNET mislabels many New Zealand Birds), then to use gemma to do the heavy lifting. If Gemma does a poor job of it, it is likely you have chosen the wrong species class.
    Keep a .md document with your mappings, BirdNET => code as below
  • edit in me.txt at line 393
    [3.744676][3.744676:744788]()
    You have access to skills /pi-ralph-wiggum to loop through the BirdNET list, and pi-heartbeat, to set a timer.
  • edit in me.txt at line 394
    [3.744789][3.744789:744934]()
    Lets manually execute this loop once, when we are happy we will design a ralph loop together to loop through remaining BirdNET classes /grill-me
  • edit in me.txt at line 396
    [3.744936][3.744936:744938]()
  • edit in me.txt at line 631
    [3.3354]
    [3.6836]
  • replacement in me.txt at line 640
    [3.8256][2.313:323]()
    os.Exit
    [3.8256]
    [3.24529]
    24/7
    ============
    rifleman--tomtit TODO
    species count calltype calltype_count
    Don't Know 22639
    Kiwi 1946 Duet 122
    Kiwi 1946 Female 126
    Kiwi 1946 Male 1698
    BlackShag 932
    comcha 485 alarm 383
    comcha 485 contact 75
    comcha 485 song 24
    comcha buzz 225
    nezbel1 455 alarm 273
    nezbel1 455 repetitive 167
    nezbel1 455 territorial 1
    identify 351
    Noise 306
    gryger1 237 song 136
    silver3 216 contact 150
    silver3 216 song 53
    saddle3 180
    nezkak1 176
    morepo2 175 morepork 1
    morepo2 175 pow 19
    morepo2 175 trill 135
    tomtit1 123 song 123
    Tree 115
    eurbla 85 contact 4
    eurbla 85 peaky 2
    eurbla 85 song 76
    Gecko 80
    nezrob3 69
    yefpar3 62
    weka1 56
    tui1 33 alarm 22
    bluduc1 31 duet 5
    bluduc1 31 rattle 4
    bluduc1 31 whistle 22
    kea1 28
    nezfan1 25 contact 21
    nezfan1 25 song 3
    pipipi1 21 song 21
    lotkoe1 16 whistle 16
    Cicada 15
    riflem1 11
    cangoo1 9
    Fly 6
    shbcuc1 6
    weta 3
    nezpig2 2 wingbeat 2
    soioys1 2
    yellow2 1 song 1
  • edit in me.txt at line 699
    [3.24586]
    cd /media/david/TwentyFourSeven/Twenty_Four_Seven
    for x in \
    R620/2024-05-06 R635/2025-02-25 R635/2026-03-23 R620/2026-03-23 \
    R778/2026-03-23 HCT_K08-24_H2/2025-06-16 R620/2025-02-25 R343/2025-02-25 \
    R306/2026-03-23 R635/2024-05-06 R287/2024-05-06 R18/2026-03-23 \
    GR5/2025-01-01 R743/2026-03-23 R287/2026-03-23 R219/2026-03-23 \
    R16/2023-12-25 J11/2026-04-06 HCT_18807_H4/2025-06-16 HCT_18103_H5/2025-06-16 \
    C03/2024-05-05 ; do
    echo ">>> $x"
    skraak calls classify --species Gecko --folder "$x"
    done
    24/7
    ======
    for x in \
    AC21/2024-01-03 AC34/2024-01-03 AC43/2024-01-03 AC83/2023-01-03 B02/2024-01-03 B22/2023-10-29 B29/2024-01-03 B29/2025-04-16 B29/2025-12-12 B40/2025-12-12 C03/2023-09-11 C03/2024-05-05 C03/2025-02-25 C03/2026-04-06 D03/2023-12-25 F09/2026-04-06 G05/2023-09-11 G05/2024-05-05 G05/2025-02-25 G05/2026-04-06 GR1/2025-01-01 GR2/2025-01-01 GR3/2025-01-01 GR4/2025-01-01 GR5/2025-01-01 GR6/2025-01-01 GR7/2025-01-01 GR8/2025-01-01 GR9/2025-01-01 H01/2023-09-11 H01/2024-05-05 H01/2025-02-25 H01/2026-04-06 H09/2023-12-25 HCT10_Kaipo/2026-05-04 HCT6_Kaipo/2026-05-04 HCT9_Kaipo/2026-05-04 HCT_18103_H5/2025-06-16 HCT_18807_H4/2025-06-16 HCT_18807_H4/2026-05-04 HCT_H13-74/2025-11-22 HCT_JOG_H7/2025-05-02 HCT_K04-14_H3/2025-06-16 HCT_K06-10_H1/2025-11-22 HCT_K08-24_H2/2025-06-16 HCT_KWR075/2025-05-04 J11/2023-09-11 J11/2024-05-05 J11/2025-02-25 J11/2026-04-06 P09/2023-09-11 P09/2023-12-25 P09/2024-05-05 P09/2025-02-25 P09/2026-04-06 R16/2023-12-25 R18/2026-03-23 R219/2026-03-23 R287/2024-05-06 R287/2025-02-25 R287/2026-03-23 R306/2025-02-25 R306/2026-03-23 R343/2024-05-06 R343/2025-02-25 R343/2026-03-23 R620/2024-05-06 R620/2025-02-25 R620/2026-03-23 R635/2024-05-06 R635/2025-02-25 R635/2026-03-23 R743/2026-03-23 R778/2024-05-06 R778/2025-02-25 R778/2025-05-23 R778/2026-03-23 ; do
    echo ">>> $x"
    skraak calls classify --species 'comcha+_' --folder "$x"
    done
    please use skill /call-classification-ollama to search all "tui1" with certainty 70 in folder
    /media/david/SSD4/Twenty_Four_Seven/R620/2024-05-06 . load some examples from /call-library for tui1+alarm. the
    task it to increase certainty of all tui alarm calls to 90, do nothing for all calls that are not alarm calls. I already started ollama with gemma
  • edit in README.md at line 55
    [3.1180601]
    [3.847]
    ./skraak calls classify --folder . --filter opensoundscape-multi-1.0 --species+_ # ie no calltype assigned, excludes species with calltype assigned