Frame ML Problem
Write ## Modeling decisions in journal/JOURNAL.md. The table
is the contract. This skill does not declare a learner and does
not evaluate one.
Human-facing prose
Details: setup-workspace references/human_facing_prose.md.
Journal rows describe this dataset. Do not name the skills
framework, the CLI, or a splitter class in the table. Questions
use data-science language — not skill ids, G-* names, or the
wrapper CLI.
Procedure
Run python -m skore_skills status. If status.setup.pending
is non-empty and status.skills.setup-ml-project is true,
load setup-ml-project and stop. Do not start this skill.
When it returns, continue. Do not load it again on this turn.
If that skill is not installed, name the pending pieces in
one line and stop. Do not invent git init, scaffold, or
env init. If status.setup.env or status.setup.workspace
is declined, stop in one line. A declined git or
editable is not asked again; continue. If data_analysis is
missing and explore-ml-data is installed, AskUserQuestion:
explore first (default) or continue from facts the user stated.
Explore loads explore-ml-data and stops. Do not invent dataset
facts. Do not ask this again once data_analysis is present
or skipped.
Run python -m skore_skills frame show. When the user is
changing a locked constraint and named one decision, do not add
--revise and do not treat proceed as the end of the turn.
Show this command once, in a fence. The only sentence for it
is: That command drops that one recorded decision. Then run
frame show with no --revise.
python -m skore_skills frame clear --cell <key>
Ask the replacement in this message, the way step 5 asks a
missing decision. Do not say that frame show will ask it.
When the named decision is the comparison metric and the
message does not state the new value, say that metric role,
prediction goal, and folds stay filled. Then name the
experiment file from the workspace state and say it still
uses the previous metric and is not run, and that the next
build or evaluate rewrites it after the table is complete
again. The last line is: The comparison
metric is MAE. Which metric replaces it? Nothing follows that
line. When the named decision is the prediction
goal and the message already states the new value, write that
value. Metric role and the comparison metric are unanswered.
Status is draft. Ask those two in this message. Do not write
Status locked. Otherwise, if the message already states the
new value, write it and ask only the decisions that are still
empty. Do not ask Modify / Keep / Stop. Do not say the new
value waits until next turn. Do not load model-ml-pipeline
or git-close on this turn. When they are changing a constraint and did not
name a decision, AskUserQuestion one pick among the filled
decisions (skip n/a) and stop. Do not also ask for a typed
answer. Do not --revise and do not edit the journal on that
turn. JSON action is authoritative. Do not invent a menu.
If the command is missing or exits without JSON, follow
Fallback below. Do not open another reference. Do not guess
candidates.
stop — say the JSON reason and stop.
ask / uncovered — follow Fallback below, including the
table written in this message and the Status rule there. Say
the reuse and change lines, and say there is no splitter
translation. Do not load build-ml-pipeline. Stop this turn.
ask / missing_keys — read each distinct reference in
questions once before asking. Do not open any other file
under references/. Ask every key in missing in one
message. For a question that has candidates, those are the
options. When candidates is absent, ask for the value the
reference describes. Draw on three sources, and only what they
actually say: the EDA report, free-form text that came with
the data if any is present (notes, a dictionary, or a README
beside the raw files), and facts the user stated. If none of
that text is present, do not invent it. When one of them
already states the fact, quote it in the question. Write every
Value the user answered in this turn. Do not rename the
Variable column. Do not stop after the first decision.
When the deployment makes other rows inapplicable, set those
rows to n/a in the same edit. Horizon, gap, and time role
are n/a unless deployment is time. Generalize-to is n/a
unless deployment is groups. A fold count of 1 is one
train/test split drawn from a single table. When the EDA
report, the text shipped with the data, or the user already
names a separate training table and test table, offer using
that split in the folds question and write predefined if
they choose it. Do not offer it otherwise. Do not write
prefit in the table. Do not type Revised on; only that
command writes that date. If any key in missing is still
unanswered, set Status to draft once any decision is filled,
ask those keys, and stop. Do not invent their values. Do not
set Status to locked. Do not ask to confirm the table. If
the write fills every required decision, set Status to
locked in that same edit, not draft. Run
python -m skore_skills frame show again in this turn. On
proceed, say the reuse and change lines, then follow step 8.
On ask / set, write Status locked only, say those lines,
run frame show again, and follow step 8. If that frame show
still returns missing_keys, the table was not complete: ask
those keys and stop, and leave Status draft.
Reuse and change lines, quoting JSON context in 2–4 lines:
these choices are reused for the rest of the experiment so
models stay comparable, and any one of them can be changed by
naming it (for example the comparison metric). Do not say
"lock" in those lines. Do not AskUserQuestion.
When the user named one decision and Status is draft, do not
treat set as accepting the table. Show the same fenced
command for that decision and stop. Do not write the new
value. Do not say any other decision was dropped. The
command's JSON list of dropped keys is the record. Status
stays draft. That command stamps Revised on; do not type
that date. The next frame show asks only keys that are still
empty or invalid.
ask / set — the table was already complete and Status is
still draft. Write Status locked only. Say the reuse and
change lines from step 5. Run frame show again and follow
step 8. Do not AskUserQuestion. The user sentence that opened
this screen is not a choice.
ask / revise is not a user question. Do not present
Modify / Keep / Stop. Ignore those choices. Drop the named
decision and ask the replacement, as in step 2.
proceed — the table is locked, and the user is not changing
a named decision. If they are, step 2 already handled it. If
translation is null, say
that this lock has no splitter translation. Do not load
build-ml-pipeline and do not return to model-ml-pipeline.
Stop. If model-ml-pipeline dispatched this turn, return to
that coordinator and stop. Do not start build, write a design
note, or run the git close from here. If no experiment script
exists and History has no running, done, or abandoned model
row, and status.skills.model-ml-pipeline is true, load that
skill and stop. Do not write a design note here. Do not
git end-turn. Otherwise run
python -m skore_skills git end-turn --stage implement. If
JSON action is invoke, load persist-ml-git only if
status.skills.persist-ml-git is true and stop. Otherwise
load triage-ml-task only if that skill is installed.
Live preview
After any turn that changes journal/JOURNAL.md (filled
decisions, Status, that command, or lock), coalesce all edits,
then, when policy.site is true and export-ml-site is
installed, run python -m skore_skills site build --if-stale
before the next question or stop. In that message, say the
website has been updated and paste the file:// line from
that output verbatim. Do not name
journal/JOURNAL.md. policy.site
null or false does not enable or build the site. A
read-only frame show with no Markdown change does not build.
Stop conditions
- Do not write Python, a pipeline, a test, or a design note.
- Do not put a class name or a constructor argument in the journal.
TimeSeriesSplit, KFold, GroupKFold, and gap= stay out of
the table.
- Do not re-ask a key that is absent from
missing.
- Do not add an option that is absent from
candidates.
- Do not open a reference the JSON did not name. When the
command is missing, follow Fallback below instead of opening
a file.
- A recorded decision changes when that command drops it, then
filling it like any missing one. A complete fill sets Status
to
locked again. Do not ask Modify / Keep / Stop.
- That command is the only journal edit that drops a recorded
decision, and only for the decision the user named. It stamps
Revised on; do not type that date. Do not rewrite
experiments/, audit/, or a report in this skill.
- After that decision is dropped, do not run an existing
experiment script. Say that it still uses the previous splitter
and metric. Show the command once, in a fence. The next build
or evaluate rewrites it after the table is complete again.
Fallback
Use this when frame show is missing, or when the problem is not
classification, regression, or multi-output regression. Do not
open any other file under references/. Do not invent the closed
candidate menu.
Tell the user, in a few lines, that the closed menu does not cover
this case, so the comparison has to be written in words. Ask one
question: what would count as a better result, and what is an honest
baseline. Draw on the EDA report, free-form text that came with
the data if any is present, and facts they stated. Quote a fact
those sources already give. If that text is absent, do not invent it.
Write the table in this same message:
- Prediction goal:
uncovered
- Metric: the comparison, in their words, or empty if they did
not state it
- Baseline note: the baseline, in their words, or empty if they
did not state it
- Every other decision row:
n/a, unless they already stated a value
Do not use placeholders. Do not leave the table for a later
message. Status is locked only when both the comparison and the
baseline are already stated. Otherwise Status is draft. The
words lock, modify, and stop stay out of the message.
Do not name a splitter class or write code. Say these choices are
reused for the rest of the experiment so models stay comparable,
and that any one of them can be changed by naming it. There is no
splitter translation.
If they say it is actually classification or regression, write that
goal instead of uncovered and stop. Do not set Status to locked.
The next frame show uses the closed menu.
If the command is missing, there is no JSON. The same write uses
the same lines. There is no splitter translation.
1---2name: frame-ml-problem3description: >4 Record the problem, deployment setting, metric, baseline, and fold count5 before any model code. Ask every missing decision in one turn, from `frame6 show`. Does not write Python, estimator hyperparameters, or splitter7 constructors.89 TRIGGER when the user asks which metric to compare on, how new rows should10 be split, which baseline to use, or says a problem constraint changed. Not11 when they ask to run evaluation or CV.1213 HOW TO USE: run `python -m skore_skills frame show`. Read each reference14 named in `questions` once, ask every key in `missing` in one message, and15 write every answered value. If a key is still unanswered, ask it and stop.16 If the write fills every required decision, set Status to `locked`, say17 those choices are reused and can be changed by name, then follow `proceed`.18 Do not ask to confirm. If that command is missing, or the recorded task is19 not classification, regression, or multi-output regression, follow Fallback20 below and do not invent the closed menu.21metadata:22 source: https://skillmd.com/skills/probabl-ai/frame-ml-problem23 modelTier: medium24---2526# Frame ML Problem2728Write `## Modeling decisions` in `journal/JOURNAL.md`. The table29is the contract. This skill does not declare a learner and does30not evaluate one.3132## Human-facing prose3334Details: `setup-workspace` `references/human_facing_prose.md`.35Journal rows describe this dataset. Do not name the skills36framework, the CLI, or a splitter class in the table. Questions37use data-science language — not skill ids, `G-*` names, or the38wrapper CLI.3940## Procedure41421. Run `python -m skore_skills status`. If `status.setup.pending`43 is non-empty and `status.skills.setup-ml-project` is true,44 load `setup-ml-project` and stop. Do not start this skill.45 When it returns, continue. Do not load it again on this turn.46 If that skill is not installed, name the pending pieces in47 one line and stop. Do not invent `git init`, scaffold, or48 `env init`. If `status.setup.env` or `status.setup.workspace`49 is `declined`, stop in one line. A declined `git` or50 `editable` is not asked again; continue. If `data_analysis` is51 `missing` and `explore-ml-data` is installed, **AskUserQuestion**:52 explore first (default) or continue from facts the user stated.53 Explore loads `explore-ml-data` and stops. Do not invent dataset54 facts. Do not ask this again once `data_analysis` is `present`55 or `skipped`.562. Run `python -m skore_skills frame show`. When the user is57 changing a locked constraint and named one decision, do not add58 `--revise` and do not treat `proceed` as the end of the turn.59 Show this command once, in a fence. The only sentence for it60 is: That command drops that one recorded decision. Then run61 `frame show` with no `--revise`.6263 ```64 python -m skore_skills frame clear --cell <key>65 ```6667 Ask the replacement in this message, the way step 5 asks a68 missing decision. Do not say that `frame show` will ask it.69 When the named decision is the comparison metric and the70 message does not state the new value, say that metric role,71 prediction goal, and folds stay filled. Then name the72 experiment file from the workspace state and say it still73 uses the previous metric and is not run, and that the next74 build or evaluate rewrites it after the table is complete75 again. The last line is: The comparison76 metric is MAE. Which metric replaces it? Nothing follows that77 line. When the named decision is the prediction78 goal and the message already states the new value, write that79 value. Metric role and the comparison metric are unanswered.80 Status is `draft`. Ask those two in this message. Do not write81 Status `locked`. Otherwise, if the message already states the82 new value, write it and ask only the decisions that are still83 empty. Do not ask Modify / Keep / Stop. Do not say the new84 value waits until next turn. Do not load `model-ml-pipeline`85 or git-close on this turn. When they are changing a constraint and did not86 name a decision, **AskUserQuestion** one pick among the filled87 decisions (skip `n/a`) and stop. Do not also ask for a typed88 answer. Do not `--revise` and do not edit the journal on that89 turn. JSON `action` is authoritative. Do not invent a menu.90 If the command is missing or exits without JSON, follow91 Fallback below. Do not open another reference. Do not guess92 candidates.933. `stop` — say the JSON `reason` and stop.944. `ask` / `uncovered` — follow Fallback below, including the95 table written in this message and the Status rule there. Say96 the reuse and change lines, and say there is no splitter97 translation. Do not load `build-ml-pipeline`. Stop this turn.985. `ask` / `missing_keys` — read each distinct `reference` in99 `questions` once before asking. Do not open any other file100 under `references/`. Ask every key in `missing` in one101 message. For a question that has `candidates`, those are the102 options. When `candidates` is absent, ask for the value the103 reference describes. Draw on three sources, and only what they104 actually say: the EDA report, free-form text that came with105 the data if any is present (notes, a dictionary, or a README106 beside the raw files), and facts the user stated. If none of107 that text is present, do not invent it. When one of them108 already states the fact, quote it in the question. Write every109 Value the user answered in this turn. Do not rename the110 Variable column. Do not stop after the first decision.111 When the deployment makes other rows inapplicable, set those112 rows to `n/a` in the same edit. Horizon, gap, and time role113 are `n/a` unless deployment is time. Generalize-to is `n/a`114 unless deployment is groups. A fold count of `1` is one115 train/test split drawn from a single table. When the EDA116 report, the text shipped with the data, or the user already117 names a separate training table and test table, offer using118 that split in the folds question and write `predefined` if119 they choose it. Do not offer it otherwise. Do not write120 `prefit` in the table. Do not type `Revised on`; only that121 command writes that date. If any key in `missing` is still122 unanswered, set Status to `draft` once any decision is filled,123 ask those keys, and stop. Do not invent their values. Do not124 set Status to `locked`. Do not ask to confirm the table. If125 the write fills every required decision, set Status to126 `locked` in that same edit, not `draft`. Run127 `python -m skore_skills frame show` again in this turn. On128 `proceed`, say the reuse and change lines, then follow step 8.129 On `ask` / `set`, write Status `locked` only, say those lines,130 run `frame show` again, and follow step 8. If that `frame show`131 still returns `missing_keys`, the table was not complete: ask132 those keys and stop, and leave Status `draft`.133 Reuse and change lines, quoting JSON `context` in 2–4 lines:134 these choices are reused for the rest of the experiment so135 models stay comparable, and any one of them can be changed by136 naming it (for example the comparison metric). Do not say137 "lock" in those lines. Do not AskUserQuestion.1386. When the user named one decision and Status is `draft`, do not139 treat `set` as accepting the table. Show the same fenced140 command for that decision and stop. Do not write the new141 value. Do not say any other decision was dropped. The142 command's JSON list of dropped keys is the record. Status143 stays `draft`. That command stamps `Revised on`; do not type144 that date. The next `frame show` asks only keys that are still145 empty or invalid.1467. `ask` / `set` — the table was already complete and Status is147 still `draft`. Write Status `locked` only. Say the reuse and148 change lines from step 5. Run `frame show` again and follow149 step 8. Do not AskUserQuestion. The user sentence that opened150 this screen is not a choice.151 `ask` / `revise` is not a user question. Do not present152 Modify / Keep / Stop. Ignore those `choices`. Drop the named153 decision and ask the replacement, as in step 2.1548. `proceed` — the table is locked, and the user is not changing155 a named decision. If they are, step 2 already handled it. If156 `translation` is null, say157 that this lock has no splitter translation. Do not load158 `build-ml-pipeline` and do not return to `model-ml-pipeline`.159 Stop. If `model-ml-pipeline` dispatched this turn, return to160 that coordinator and stop. Do not start build, write a design161 note, or run the git close from here. If no experiment script162 exists and History has no running, done, or abandoned model163 row, and `status.skills.model-ml-pipeline` is true, load that164 skill and stop. Do not write a design note here. Do not165 `git end-turn`. Otherwise run166 `python -m skore_skills git end-turn --stage implement`. If167 JSON `action` is `invoke`, load `persist-ml-git` only if168 `status.skills.persist-ml-git` is true and stop. Otherwise169 load `triage-ml-task` only if that skill is installed.170171## Live preview172173After any turn that changes `journal/JOURNAL.md` (filled174decisions, Status, that command, or lock), coalesce all edits,175then, when `policy.site` is `true` and `export-ml-site` is176installed, run `python -m skore_skills site build --if-stale`177before the next question or stop. In that message, say the178website has been updated and paste the `file://` line from179that output verbatim. Do not name180`journal/JOURNAL.md`. `policy.site`181`null` or `false` does not enable or build the site. A182read-only `frame show` with no Markdown change does not build.183184## Stop conditions185186- Do not write Python, a pipeline, a test, or a design note.187- Do not put a class name or a constructor argument in the journal.188 `TimeSeriesSplit`, `KFold`, `GroupKFold`, and `gap=` stay out of189 the table.190- Do not re-ask a key that is absent from `missing`.191- Do not add an option that is absent from `candidates`.192- Do not open a reference the JSON did not name. When the193 command is missing, follow Fallback below instead of opening194 a file.195- A recorded decision changes when that command drops it, then196 filling it like any missing one. A complete fill sets Status197 to `locked` again. Do not ask Modify / Keep / Stop.198- That command is the only journal edit that drops a recorded199 decision, and only for the decision the user named. It stamps200 `Revised on`; do not type that date. Do not rewrite201 `experiments/`, `audit/`, or a report in this skill.202- After that decision is dropped, do not run an existing203 experiment script. Say that it still uses the previous splitter204 and metric. Show the command once, in a fence. The next build205 or evaluate rewrites it after the table is complete again.206207## Fallback208209Use this when `frame show` is missing, or when the problem is not210classification, regression, or multi-output regression. Do not211open any other file under `references/`. Do not invent the closed212candidate menu.213214Tell the user, in a few lines, that the closed menu does not cover215this case, so the comparison has to be written in words. Ask one216question: what would count as a better result, and what is an honest217baseline. Draw on the EDA report, free-form text that came with218the data if any is present, and facts they stated. Quote a fact219those sources already give. If that text is absent, do not invent it.220221Write the table in this same message:222223- Prediction goal: `uncovered`224- Metric: the comparison, in their words, or empty if they did225 not state it226- Baseline note: the baseline, in their words, or empty if they227 did not state it228- Every other decision row: `n/a`, unless they already stated a value229230Do not use placeholders. Do not leave the table for a later231message. Status is `locked` only when both the comparison and the232baseline are already stated. Otherwise Status is `draft`. The233words lock, modify, and stop stay out of the message.234235Do not name a splitter class or write code. Say these choices are236reused for the rest of the experiment so models stay comparable,237and that any one of them can be changed by naming it. There is no238splitter translation.239240If they say it is actually classification or regression, write that241goal instead of `uncovered` and stop. Do not set Status to `locked`.242The next `frame show` uses the closed menu.243244If the command is missing, there is no JSON. The same write uses245the same lines. There is no splitter translation.