Backend of Quantist.io: Capturing The Silhouette of Data

Democratize the information of Indonesia Stocks and more to investors from retail to institution.

Consists of high-end analysis tools based on data with top-down analysis:

• Conservation of money (forex/crypto - bond - stocks)
• External factor (commodities, national/international issues, sentiments)
• Fundamental: business performance, corporate action
• Transaction: foreign flow, broker summary clustering, done trade, 5% shares holder, share holder distribution
• Behaviour: supply & demand, momentum, trend, time

For more information, you can find us at:

• Website
• Email

Volume-profile screener criteria (frozen contract)

The foreign and broker volume-profile screeners accept these screener_vprofile_criteria query values:

• vprofile_inside — membership only: a recent close touched the selected profile zone.
• vprofile_breakout — exact role=resistance, behavior=breakout_up: approach from below, touch, then close above.
• vprofile_breakdown — exact role=support, behavior=breakdown: approach from above, touch, then close below.
• vprofile_support_bounce — exact role=support, behavior=rejection: approach from above, touch, then close above.
• vprofile_resistance_rejection — exact role=resistance, behavior=rejection: approach from below, touch, then close below.

The directional criteria require their exact role and behavior pair; a role alone, membership alone, or a close still inside the zone is not selected.

Selected-zone annotations also include vprofile_zone_strength (absolute node flow relative to the strongest node), vprofile_zone_prominence (raw peak/valley prominence relative to the strongest absolute histogram flow), vprofile_zone_flow_share (absolute node flow divided by total absolute profile flow), and vprofile_event_date (latest close date as YYYY-MM-DD). These are JSON-safe scalars; all four are None when no profile zone is selected. Existing zone levels, role, behavior, distance, and touch count remain unchanged.

Web chart contract (/web-api/v1/chart/{code})

web_contract.py is the single source of truth for the semantic chart payload the web frontend consumes. It is versioned (schema_version, currently 1.0) and deliberately independent of /whaleanalysis, which keeps returning rendered Plotly figures for Telegram and the existing public API. Nothing in the web contract is coupled to a chart library.

Sections: instrument, period, summary, price, indicators, profile_zones, holding_composition, annotations, meta.

Guarantees the frontend relies on, pinned by test_web_contract.py:

• Numbers, never formatted strings. proportion is 0.3342, not "33.42%"; money_flow is -1.0e12, not "-1.00T". Ratios stay ratios (correlations -1..1, proportions 0..1); money and price stay in their native units.
• null for missing. NaN and infinity are normalized to null; a column the source never produced comes back as an all-null series of the right length.
• One date axis. Every price.* and indicators.* list is aligned index-for-index with price.dates, so the frontend never re-joins on date.
• ISO dates, explicit timestamp. Dates are YYYY-MM-DD; meta.generated_at is a timezone-aware ISO 8601 datetime.
• Explicit profile-zone boundaries. Each zone carries low, high, mid, net_value and is_peak; a boundary is never implied by a mid point.

COMPOSITE and IHSG resolve to the same instrument (backend code composite, display_name "IHSG / COMPOSITE"). Resolution is case- and whitespace-insensitive, and an empty code falls back to composite.

Authentication is unchanged: the route is mounted behind the same X-API-KEY dependency as every other router, so a browser is never the direct caller. The frontend repo's Cloudflare worker is the intended caller and holds the key as a worker secret.

Screener ordering

Money-flow accumulated results rank by total flow descending; distributed results rank ascending, so the strongest negative flow is first. Ties use ascending stock code. VWAP rally ranks by the largest positive close-to-VWAP gap. VWAP around ranks by the smallest absolute gap. VWAP breakout and breakdown rank by the freshest valid cross first, then the smallest distance from VWAP, then price follow-through. Flow is not used for VWAP ordering. Volume-profile rankings use prominence, then strength, then flow in the event direction (inside/upward descending, downward ascending), with ascending stock code as the final tie-break. Volume-profile annotations are computed for every criterion member before the requested limit is applied.

Copyright (c) 2023 Quantist.io. All rights reserved. This works, including all modifications made by any third party, is the property of Quantist.io and is protected by copyright law.