Pull-everything survey: 76 row-level action exporters cataloged, ~14 consumed #7
Labels
No labels
burndown-2026-06
burndown-2026-08
headless
autonomy
async-consult
autonomy
epic
autonomy
headless
autonomy
live-collab
c#
priority
P0
priority
P1
priority
P2
priority
P3
priority
P4
role/advocate
role/director
role/exec
role/frontend
role/gamedev
role/human
role/platform
role/qa
role/science
role/sysadmin
state
ambient
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
coilyco-gaming/eco-app#7
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
While answering "is trades data only a counter?" the agent enumerated
/datasets/flatlist: 205 datasets, 76 of them row-level action exporters (IsAction: true, served as CSV via/api/v1/exporter/actions?actionName=...). The MCP tools currently consume only ~14 of them as aggregate/datasets/getseries plus the 4 crafting actions.Standing direction (now an AGENTS.md rule): pull everything remotely interesting - code is cheap, the server is colocated, CPU is fine. This issue is the umbrella tracker; each dataset that gets pulled and turns out messy gets its own cleanup issue.
High-interest candidates from the catalog, roughly grouped:
Shared cleanup dependencies: id-to-name join (#5), undecoded enums, per-action column-shape quirks (misalignment, #5), Time-to-day conversion.
Full 76-name catalog captured 2026-06-12 from the live cycle-13 server.
Full populated-dataset survey committed: docs/dataset-survey-cycle13.md. 192 of 205 datasets are live: 68 action exporters (row-level CSVs) + 124 daily series. Headline volumes: PayTax 6,909 rows, TransferMoney 5,979, ReceiveGovernmentFunds 5,436, ClaimOrUnclaimProperty 4,825, plus 94 species population series. Kai reviews the list and picks pulls; Eco source access at ~/projects/StrangeLoopGames/Eco is now available for probing beyond the exporter surface.
🔎 ward agent advisor
ward agent advisor --driver clauderan a one-shot standard research pass on this question:Grounded in the live catalog (
docs/datasets/, cycle-13) and the actual consumers insrc/eco_mcp_app/(server.pyECONOMY_DATASETS,crafting.pyCRAFT_ACTION_TYPES). Here is the ranked mapping.What "consumed" actually means today
Two distinct consumption modes are already in the tree, and the distinction drives the whole plan:
/api/v1/exporter/actions) - only 4 actions, all incrafting.py:ItemCraftedAction,HarvestOrHunt,ChopTree,DigOrMine./datasets/get) - the 14 inECONOMY_DATASETS: the loan quartet,PayWages,PayRentOrMoveInFee,PostedContract,CompletedContract,FailedContract,PropertyTransfer,ReputationTransfer,TransferMoney,PayTax,ReceiveGovernmentFunds.Key consequence: those 14 are consumed as counters only. Their row-level exporters (who paid whom, how much, when) are still unconsumed. For #12 the highest-value move is not "pull new actions" but "pull the row-level form of actions we already count." That reframes the leverage math below.
Per-ticket exporter lists (the decided lists)
#6 trades ledger - the named, narrowest, highest-signal target.
CurrencyTrade(the ledger's spine;BoughtOrSoldenum 32/33 is the #6 decode blocker)BarterTrade(non-currency exchange; completes "all trades," not just money trades)TradesInLast7Daysseries (latest ~1.8M currency) as the sanity cross-check that row sums reconcile.#12 economic health dashboard - upgrade counters to distributions, then add the money-supply side.
PayTax(6,909 rows),TransferMoney(5,979),ReceiveGovernmentFunds(5,436),PayWages,PayRentOrMoveInFee- these turn today's single-number KPIs into per-citizen distributions (tax burden, wage dispersion, a real Gini).MintCurrency,CreateCurrency(inflation / new-currency events the counter set has no view of).CurrencyTrade/BarterTradeshared with #6 (velocity + market depth).ClaimOrUnclaimProperty(4,825 rows) +PropertyTransferrow-level for the wealth-in-land dimension.#17 crafting activity atlas - the 4 core actions ship; enrich toward the full production pipeline.
CreateWorkOrder,LaborWorkOrderAction,PostedWorkParty,JoinedWorkParty,WorkedForWorkParty,CompletedWorkParty.PostedContract,CompletedContract,FailedContract(currently counters only in #12's set - shared).GainProfession,GainSpecialty,SpecialtyLevelUp,CompleteClass,EnrollAction- also the natural/jobshistory enrichment.#20 biodiversity drift - correlate the 94 flora/fauna population series (already live via
/datasets/get) against the player actions that move them.HarvestOrHunt(2,471) andChopTree(1,076) - already pulled by crafting, re-frame the same rows biodiversity-first.PlantSeeds(690),PlowField(640),FertilizeAction(5),CreateTreeDebris(4,326).PolluteAir(2,109) row-level, cross-referenced with the climate CO2 series.Highest-leverage consumption order
Ranked by (tickets unlocked) x (signal density) / (cleanup cost), and sequenced so shared cleanup work is paid once and reused.
1.
CurrencyTrade+BarterTrade- do first.Unlocks #6 outright and is a core input to #12 (double-duty). It is the ticket with a name attached, the data is high-volume and high-signal, and it forces the two cleanup dependencies (#6 enum decode, #5 id-to-name join) that everything downstream also needs. Pay that tax here where a shipping feature justifies it.
2. Row-level upgrade of the counted-14 for #12 -
PayTax,TransferMoney,ReceiveGovernmentFunds,PayWages,PayRentOrMoveInFee, thenMintCurrency/CreateCurrency.Second because it reuses the #5 join stood up in step 1 and turns the existing economy dashboard from counters into distributions - the biggest quality jump per row pulled, on a surface that already exists. No new enum work (money amounts, not coded fields).
3. Biodiversity action set for #20 -
PlantSeeds,PlowField,FertilizeAction,CreateTreeDebris,PolluteAir, plus re-framedHarvestOrHunt/ChopTree.Third because it is the cheapest cleanup profile (mostly
Time+ quantity, no enums, less blocked by #5 since the payoff is action-vs-population correlation, not per-citizen identity) and the 94 population series are already consumable - so this is "join new action rows to data we already have," a fast win once the two economic tickets have absorbed the shared plumbing.4. Work / production + progression for #17 -
CreateWorkOrder,LaborWorkOrderAction, the work-party family, contracts row-level, then the skill/progression actions.Last of the four because it enriches an already-shipping feature (lowest marginal user value), carries the heaviest column-shape risk (work orders and labor have the most per-action schema quirks flagged in #5), and the progression subset overlaps
/jobs- worth splitting into its own run once the atlas core is stable.Shared cleanup prerequisites (do once, in step 1)
CurrencyTrade.BoughtOrSold32/33 and any sibling coded fields; step 1 territory.Time-to-day conversion - universal (seconds-since-cycle-start), trivial, factor into the shared exporter helper so every later pull gets it free.Net: steps 1-2 front-load essentially all the reusable cleanup, so #20 and #17 runs start with the plumbing already paid and just wire their exporter lists.
Out of scope note:
ChatSent(full chat log),Vote/StartElection/civics, andPlay/FirstLoginsocial actions map to none of these four tickets. They are worth their own feature tickets before consumption - do not fold them into these runs.Researched and posted automatically by
ward agent advisor --driver claude(ward#179). This is one-shot research, not a carried change - verify before acting on it.— Claude (she/her), via
ward agentThe survey and its prioritization are complete, so this no longer needs a standing Kai selection checkpoint.
docs/datasets/.Keeping #7 open would turn a completed survey into an indefinite product backlog. New dataset pulls should be bounded source issues. Closing the completed survey.