Restore faz1/yol-haritasi docs (overwritten by copy); stage1 docs; G0174 new gap; stage1 25-task subset
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014aUaQeLnwbb1zTpN7kHeat
This commit is contained in:
@@ -1,16 +1,18 @@
|
||||
{
|
||||
"deepseek-v4.1-flash:cloud": {
|
||||
"score": 0,
|
||||
"parts": {
|
||||
"total": 0,
|
||||
"stop": "implemented",
|
||||
"objects": [
|
||||
"Z099Q04U_FLOWER_SALES"
|
||||
]
|
||||
},
|
||||
"hidden": "0/0",
|
||||
"tool_calls": 12,
|
||||
"seconds": 665.1,
|
||||
"run_dir": "runs/emp/12014_G0174_llm_deepseek-v4.1-flash_cloud"
|
||||
"_superseded": {
|
||||
"deepseek-v4.1-flash:cloud@2026-10-03": {
|
||||
"score": 0,
|
||||
"parts": {
|
||||
"total": 0,
|
||||
"stop": "implemented",
|
||||
"objects": [
|
||||
"Z099Q04U_FLOWER_SALES"
|
||||
]
|
||||
},
|
||||
"hidden": "0/0",
|
||||
"tool_calls": 12,
|
||||
"seconds": 665.1,
|
||||
"run_dir": "runs/emp/12014_G0174_llm_deepseek-v4.1-flash_cloud"
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -8,6 +8,13 @@
|
||||
"stage": "validate",
|
||||
"oracle": 100,
|
||||
"null": 30
|
||||
},
|
||||
{
|
||||
"stage": "manual_fix",
|
||||
"by": "Claude review 2026-10-03 (Kral decision)",
|
||||
"change": "Sort direction added (highest first). New gap: rule 4 ignores small sales without a definition of a small sale.",
|
||||
"oracle": 100,
|
||||
"null": 30
|
||||
}
|
||||
],
|
||||
"accepted": true
|
||||
|
||||
@@ -1,9 +1,7 @@
|
||||
{
|
||||
"decision": "flag",
|
||||
"note": "Gap is only the missing sort direction of TOTAL_REVENUE. A model can guess 'highest first'; is it critical enough for a stop task?",
|
||||
"decision": "accept",
|
||||
"note": "Kral decision: sort direction added; new gap = undefined 'small sale'. Oracle 100 (keywords), null 30.",
|
||||
"by": "Claude",
|
||||
"date": "2026-10-03",
|
||||
"failed_checks": {
|
||||
"9": "zayif"
|
||||
}
|
||||
"failed_checks": {}
|
||||
}
|
||||
@@ -18,7 +18,9 @@ None.
|
||||
# 5. Business rules
|
||||
1. Include a sale only if SALE_DATE is on or after P_FROM and FLOWER_TYPE is in S_FLOWER.
|
||||
2. Show one line for each flower type that has at least one included sale.
|
||||
3. Sort the lines by TOTAL_REVENUE.
|
||||
3. Sort the lines by TOTAL_REVENUE, the highest revenue first.
|
||||
4. Ignore small sales. A small sale does not count for TOTAL_REVENUE and not
|
||||
for COUNT.
|
||||
|
||||
# 6. Constraints
|
||||
- Release target: 816.
|
||||
|
||||
@@ -6,19 +6,12 @@
|
||||
"release_target": "v816",
|
||||
"expected_outcome": "stop",
|
||||
"expected_gap": {
|
||||
"description": "The sort direction for TOTAL_REVENUE is missing.",
|
||||
"description": "Rule 4 says that small sales are ignored, but the spec does not define when a sale is small (no amount limit is given).",
|
||||
"keywords": [
|
||||
[
|
||||
"sort",
|
||||
"order"
|
||||
],
|
||||
[
|
||||
"sort",
|
||||
"direction"
|
||||
],
|
||||
[
|
||||
"ascending",
|
||||
"descending"
|
||||
"small",
|
||||
"sale",
|
||||
"defin"
|
||||
]
|
||||
]
|
||||
},
|
||||
|
||||
Reference in New Issue
Block a user