The user pointed out that the only genuinely per-user AI surface is
portfolio analysis. The strategic log AND the email digest are both
shared cycles — generated once per cycle, consumed by many users.
For the digest, this means:
- _generate_variants still produces one English variant per tone (as
today, unchanged)
- A new helper translates each variant once per active non-en lang in
parallel via asyncio.gather, producing a {(tone, lang): content}
table for the duration of the job run
- The per-user send loop selects (user.digest_tone, user.lang),
falling back to the English variant of the same tone on miss
Translation count per run = tones × non-en active langs = 3 today.
100 Italian users no longer mean 100 translation calls.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
11 TDD-style tasks: i18n service, translation helper, model + migration,
ai_log_job translation fan-out, per-user surfaces (analyse, digest),
localized /log endpoint, PATCH /api/settings/language, dropdown UI, and
final regression + manual smoke.
Per-user surfaces append "Respond in Italian." to the system prompt
(one extra line, no extra LLM call). The strategic log is generated in
English, then fanned out to translate() per active non-en language in
parallel via asyncio.gather. The /log endpoint serves the matching
translation row when present, English fallback otherwise.
Translation uses the default call_llm provider chain — no separate
cheap-model carve-out needed at DeepSeek's $0.28/M output pricing.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Translate for any user with lang='it' regardless of paid/free status.
Italian + UK are the first markets, so IT availability is part of the
public-facing experience — a free-tier visitor needs to see the AI in
Italian to convert. At ~$0.005/day total cost the gating isn't worth
the savings.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Hybrid model: per-user surfaces (analyse, digest, chat) generated
directly in the target language via a "Respond in Italian" clause
appended to the system prompt. Shared content (strategic log)
generated in English as today, then post-translated and cached per
language in a new strategic_log_translations table. Translation calls
fan out in parallel with asyncio.gather so total job latency stays
bounded by max(single call).
No separate translation-model setting — DeepSeek-4-flash at $0.28/M
output is cheap enough that the routine cost is noise (~$0.005/day
with Italian only at 24 logs/day).
Users.lang VARCHAR(8) DEFAULT 'en'. Settings dropdown lists all four
options but ES/FR/DE are disabled UI-side and rejected server-side
against an ACTIVE_LANGUAGES allowlist — flipping them on later is a
one-line constant change.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Delete app/templates/upload.html. The /upload route redirected to
/settings#import (302) and never rendered this template; the file
was carrying stale Trading-212-only copy.
- Landing + pricing pages: replace "Trading 212 today, more brokers
planned" with "Trading 212 natively, other formats auto-detected"
to reflect the LLM-fallback parser that's been live for a few days.
The /upload redirect route in app/routers/pages.py stays — it remains
a useful bookmark-forwarder for users with old links.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
A small italic muted line beneath the form explaining the controls:
"Type a symbol, then quantity and cost — or use the calendar to fill
cost from a buy date — then [+] to add. [×] next to an existing row
removes it."
Only renders while the composer itself is visible (i.e. in edit mode),
so it doesn't clutter the dashboard at rest.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Two related polishes:
- The add form was auto-shown by the empty-state path so brand-new
users would see something to act on. That conflicts with the user's
preference for "Edit always toggles the form, no other path." The
empty state now shows guidance copy ("click edit to add one")
instead. exitEditMode always hides the form too.
- The submit "add" word-button is replaced by a square accent-bordered
+ glyph (26×26). Matches the visual weight of the calendar ghost
next to it but stays in the accent colour so it reads as primary.
Adds a tiny active-state scale tick for tactile feedback.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Composer had zero horizontal padding so the leading `$` prompt was
flush with the panel border. Match the panel-header's 12px horizontal
inset so the form sits inside the panel's content gutter.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Subtitle was technical noise that didn't earn its space in the header.
Title alone reads cleaner. Kept the scoped panel-header layout override
in cassandra.css since it's harmless and future-proof against re-adding
header children.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
A single-quoted string literal "couldn't validate" was breaking the
parse because the apostrophe wasn't escaped. The page logged a syntax
error and none of the edit-mode JS ran. Backslash-escape it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Two interlocking bugs surfaced after the design pass:
1. CSS `display: inline-flex` on .pf-edit-btn/.pf-done-btn overrode the
UA's `[hidden] { display: none }`, so the JS toggling `editBtn.hidden`
had no visual effect — both buttons rendered side by side.
2. portfolio.js's empty-state path sets `form.hidden = false` but the
populated-portfolio render path only removed the `pf-empty` class; it
never reset `form.hidden = true`. So once a user went through the
empty state, the add form stuck around — leaving the Add button
visible on a populated dashboard.
Fixes are surgical: add an explicit `[hidden]` rule for the two
header pills, and re-hide the form in `renderPanel` unless edit mode
is currently active (so we don't yank the form out from under an
edit-in-progress).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The previous CSS used invented variable names (--neu-dim, --err, --ok)
that don't exist in the project's design system; the form fell back to
hardcoded hex values and looked disconnected from the rest of the site.
Rebuilt against the real tokens (--border, --dim, --muted, --positive,
--negative, --warning, --accent) and the mono-first 'geopolitical-
terminal aesthetic' the rest of the dashboard uses:
$ ticker ✓ 172.40 USD │ qty @ cost USD 📅 add
────
- No boxed-form chrome. A dashed bottom rule separates the composer
from the table below.
- Inputs lose their card-style boxes; they're underline-only with a
faint accent wash on focus — feels like editing a command line.
- '$' prompt marker, '│' divider, '@' between qty and cost give the
row a terminal grammar without being twee.
- Submit is a ghost pill in the accent colour; lights to solid only
when enabled.
- All controls now respond correctly to the light/dark theme toggle.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Replace the multi-row wizard-style form (Ticker / Qty on row 1, mode
radios on row 2, Date+Cost on row 3) with a single horizontal strip
that sits unobtrusively above the portfolio table. The radio toggle is
gone; a small calendar icon next to the Cost input pops out a date
picker that auto-fills cost on selection and then hides itself.
Same input IDs, so the existing validate/Add/× handlers work unchanged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
12 TDD-style tasks: two backend endpoints (validate + historical),
router registration, dashboard markup, and five JS slices building the
edit-mode behaviour (toggle → ticker validate → Add → date-mode →
delete via delegation). CSS pass and final manual smoke close it out.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Dashboard-native edit mode: EDIT button toggles in-place editing; the
add-position form has on-blur ticker validation against a new paid
endpoint, qty input, and an avg-cost / bought-on-date toggle. Only
avg_cost + qty are persisted to localStorage (no acquisition date,
no server-side holdings). Empty state replaces "Import a CSV" with
the inline form so brand-new users can act without leaving the page.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The route's resolve-slice loop is T212-specific — it looks tickers up
against the InstrumentMap, which only has T212's universe. For the LLM
path the ticker is already Yahoo-ready (e.g. VOD.L, ASML.AS), so
sending it through resolve_slice produced spurious "could not be
resolved" warnings and dropped the positions.
Fix: ParsedPie gains a ``tickers_resolved`` flag (default False for
T212 backward-compat); _apply_mapping in the LLM path sets it True
and also extracts currency from the LLM-mapped currency_col into a
new ``ParsedPosition.currency`` field. The route branches on the flag:
LLM-path positions are kept verbatim with a best-effort InstrumentMap
lookup for nicer name/currency overrides, never dropped.
Integration test tightened to assert all 5 IBKR fixture positions
round-trip with the right currencies (USD / GBP / EUR).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- portfolio.js empty-state CTA: "Import a T212 CSV" → "Import a portfolio CSV"
- settings.html lede: lead with broker-agnostic copy; relegate the T212
export path to a smaller secondary line.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Section heading drops "Trading 212"; drop-zone label and hint mention
the auto-detect path; the help-paragraph opens conditionally with
"If you use Trading 212" so non-T212 users don't feel like outsiders.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Heuristic refined from the plan draft: candidate header rows must be
followed by a row containing at least one numeric token. Without this,
IBKR-style multi-line preambles (all-text rows before the real header)
would be mistaken for the header at preamble=0.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Drop first_seen_user_id; sample is anonymous by construction
- Rename sample_dummy → sample_row, store the upload's first real data
row verbatim (one row, no totals, no other positions, no link to a
user). Narrow, deliberate exception to the "no holdings persisted"
invariant — gives the operator material for hand-writing future
native parsers.
- Drop the cache self-heal behaviour; operator owns eviction. Reinforce
the non-goal of auto-promoting learned formats to code.
Transparent fallback after parse_t212_csv: LLM extracts a column-mapping
(not the data), result is cached globally by header fingerprint, replay
is deterministic Python. Stored dummy contains headers + synthetic row
only — no user holdings ever persisted.
The "Phase D.1/D.2/D.3" comment scaffolding and the "Paddle webhook
will fill this in" references became actively misleading after D.3
landed — anyone reading the code would think referral conversion was
still pending. Also corrects a stale "Paddle" reference to "Stripe"
(we never shipped Paddle; ended up on Stripe after the Paddle → Polar
→ Stripe MoR onboarding pivot).
Pure docstring sweep, no behaviour change.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The referral feature was half-built: codes captured, banner shown,
counts displayed — but no money flowed when a referred user paid.
The Settings page hard-coded "— (D.3)" for Active credits and the
marketing copy promised "50% off for 3 months" with nothing behind it.
Closing the loop:
- New `convert_referral(session, user)` in referral_service.py looks
up the user's Referral row, stamps `converted_at` + `credited_at`,
and extends `credit_until` by 45 days on BOTH the buyer and the
referrer. Idempotent — replayed webhooks and renewals are no-ops.
Stacks correctly when the user already has a credit window running
(anchors at max(now, current_credit_until) like cli.grant_credit).
- Stripe webhook wires this into `_grant_paid`. A captured
`first_paid_transition = user.tier != "paid"` gate avoids the DB
lookup on every renewal event; convert_referral's own idempotency
is the second line of defence.
- `_grant_paid` now takes `session` as its first positional arg so
the conversion runs inside the same transaction as the tier flip
and audit-row write. A mid-flight failure rolls everything back
together — no partial state.
- Settings page replaces the "— (D.3)" placeholder with the live
count of conversions still inside their 45-day credit window, plus
a "+N days on your account" hint when the user has any credit of
their own (referrer bonus, admin grant, or future refund-as-credit).
- Marketing copy on pricing.html + settings.html switches from "50%
off for 3 months" to "45 days of paid access" — same economic value,
honest about the actual mechanism (full free access rather than
discounted billing).
Credit-amount rationale: 50% × 3 months ≈ 1.5 months of free
service ≈ 45 days. Pure-credit delivery is processor-agnostic, needs
no Stripe coupon plumbing, and stacks cleanly across referrals.
7 new tests in test_referral_conversion.py cover the happy path,
idempotency, no-referral no-op, credit stacking, deleted-referrer
survival, end-to-end webhook → credit landing, and the renewal-event
no-double-credit guarantee.
Also bundled: the Restore-button class fix from earlier
(portfolio.js — the cloud-restore "Restore" submit was unstyled and
picked up browser defaults; now uses .settings-btn like the rest of
the action-button family).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Settings page tidy-up driven by user feedback that it had grown too busy:
- Each section (Import, Invite, Email digests, Cloud sync) is now a
native <details>/<summary> accordion. Import stays open by default
because /settings#import is the deep-link target from the dashboard
CTA; the others collapse so the page lands quiet.
- Manage subscription is a right-aligned gear-icon button instead of
a rectangular text button — the descriptive copy moves into the
tooltip. Frees up the Tier row of visual weight.
Auth + modal inputs were too small (verify code box, portfolio restore
PIN): the auth-card selector now covers text inputs as well, and a new
.modal-input class standardises 16px / 12px-padding fields used in the
cloud-sync enable modal and the portfolio restore prompt.
The verify page no longer carries the "Email me the digest" checkbox —
it was misleading on repeat logins (server-side it only applied on
first sign-up but rendered every time). Default-opt-in lives in the
User row at creation; per-user changes happen on /settings. First
successful verify now triggers a one-shot welcome email explaining the
digest cadence and pointing at /settings for opt-out; SMTP failure is
logged but does not block the login.
Tests rewritten to cover the new welcome-email path:
- first login sends exactly one welcome email
- returning user gets none
- SMTP failure does not break the redirect
- regression guard: returning user who opted out stays opted out
Also lands the paddle merchant-summary doc that was written earlier
during the Paddle → Polar → Stripe onboarding pivot.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Bundles three related pieces that came out of the operator's first
end-to-end test of the paid flow:
1. Manage subscription button on /settings (paid users with a real
Stripe sub — i.e. not credit-granted access). POSTs to the existing
/api/stripe/portal endpoint; Stripe-hosted customer portal handles
card updates, cancellation, monthly↔annual switch, invoice history.
Replaces the stale "Paid features unlock with Paddle (D.3) or
invite credits" hint for free users with a live link to /pricing.
2. Per-cadence cooling-off treatment:
- **Annual £70**: 14-day free trial via
subscription_data.trial_period_days=14. No money moves during
the trial, so the CCR 2013 14-day refund question doesn't arise
(nothing paid = nothing to refund). Card is still required at
checkout so Stripe can charge on day 15.
- **Monthly £7**: bills immediately. A 14-day trial there would
give away ~50% of cycle one. Instead, /pricing now carries a
required tick-box above the Subscribe buttons (subscribe stays
disabled until checked) — by ticking, the user expressly
consents to begin performance immediately and acknowledges that
this extinguishes their statutory 14-day right under Reg 36
CCR 2013. Consent collected on our own page (not via Stripe's
account-wide consent_collection.terms_of_service) so each
product can keep its own Terms URL as we add more.
3. T&C §6 clause 1 split into 1a (annual / trial substitute) +
1b (monthly / Reg 36 waiver via on-page tick-box). Clause 2
(post-cooling-off cancellation) unchanged.
Settings page shows "Free trial — N days remaining" while the
sub is in `trialing` status, falling back to "Paid subscription
active." once it transitions to active. Countdown is computed
server-side from User.stripe_trial_end_at (new column, migration
0020) populated by the subscription.created/updated webhook from
the Stripe trial_end timestamp; cleared on the trialing→active
transition and on revoke.
Drive-by: fixed a structlog kwarg-name collision on
`log.warning(..., event=event_type, ...)` in both polar_webhook.py
and stripe_billing.py — `event` is structlog's positional event
name and "got multiple values" crashed the user-not-found log
path. Renamed to `event_type=` everywhere it appeared. Caught by
the new trialing-stores-trial-end test.
Tests
- 4 new in test_stripe_billing.py covering monthly (no trial, no
consent_collection), annual (trial, no consent), trialing stores
trial_end, trialing→active clears trial_end.
- 1 existing test renamed + reworked for the consent split.
- Full suite: 224 passed, 5 skipped.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Bug: the per-browser pie was stored under a single global key
(`cassandra.pie`) with no per-user scope. If User A uploaded a
portfolio and User B then signed in on the same browser, User B saw
User A's holdings — portfolio.js read straight from localStorage on
hydration with no check that the data belonged to the current session.
This was not a server-side leak: the session cookie was correct, no
API returned User A's data to User B. The stale browser state was the
sole vector. Reported by the operator while testing the paid-checkout
flow with a second account on the same browser.
Fix — defense in depth, two layers:
1. base.html now stamps cu.user.id into localStorage as
`cassandra.user_id` on every authenticated page load. If the
previous stamp doesn't match the current user, wipe localStorage
(preserving only `cassandra.theme`, which is cosmetic) and
sessionStorage before any other script runs. This catches:
- the reported scenario (User A logs out, User B logs in)
- any case where logout missed the wipe (JS disabled, browser
killed before the redirect ran)
- cookie-revocation / session-rotation edge cases where the
server-side identity changes without an explicit logout
2. /logout no longer returns a bare 303; it returns a small HTML
page that actively wipes per-user localStorage + sessionStorage
client-side (theme preserved), then redirects to /login. A
meta-refresh covers the no-JS case (the cookie deletion is
still server-side, so security is preserved either way).
Behaviour for the legitimate case (same user logs out + back in)
is unchanged: their localStorage data survives because the
mismatch check sees the same user_id and doesn't fire — the
logout wipe runs but they re-stamp + re-upload only the
cassandra.user_id and a fresh pie cycle if they choose to upload.
Suite: 221 passed, 5 skipped, 0 failed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
