EGTS/Guide
Search current listings, ISIN, BSE codes...⌘K
Sign In
In-Product Manual & Knowledge BaseVIEW ONLY

EGTS Product & Operator Guide

Authoritative documentation for all 29 product areas: workflows, models, alert delivery, system architecture, and safety boundaries.

(All 29 sections, 41 alert types, operational glossary & synonyms)
Press Esc to clear
Quick Topics:

1. Getting Started

Getting Started & WorkflowVIEW ONLY
What It Is

An introduction to the EGTS market intelligence platform, designed for institutional-grade screening and research across Indian equities without broker order execution.

Why It Exists

Provides a structured starting point so new operators and analysts can navigate the platform safely and effectively.

When to Use:Use when onboarding to EGTS or orienting yourself with core analytical workflows.
How to Use:Begin with Discover for precomputed opportunities, move to Scanner for custom criteria, inspect Stock Detail for individual setups, and monitor Alerts.
Expected Outcome:Clear understanding of platform capabilities, analytical boundaries, and navigation structure.
Important Operational Invariants & Rules:
  • EGTS is strictly an analytical intelligence platform β€” NO broker order execution takes place.
  • Model Setup Scores rank candidates by indicator confluence; they are NOT guaranteed win probabilities.
  • Single-writer architecture: the Control Plane handles all autonomous data generation.
View-Only Scope:All stock analytics, rankings, historical charts, and publication manifests are strictly view-only.
Safe to Change:Personal watchlists, custom alert rules, and recipient notification preferences.
Never Touch:Production parquet files, publication manifests, or lock files manually.
βœ“ Healthy Normal State:Green status badges across Platform EOD session, Control Plane heartbeat, and clean publication gate.
⚠️ When Something Is Wrong:Stale data banners or red failure alerts in system cockpit.
Example: An analyst opens EGTS at 08:45 IST, checks Discover for Swing candidates, and saves 3 top setups to their watchlist.
Next: Proceed to 'Daily Workflow' to understand daily operational routines.Open Discover Page
#getting started#overview#introduction#welcome#basics#how to start#first steps

2. Daily Workflow

Getting Started & WorkflowVIEW ONLY
What It Is

A standardized chronological 4-stage daily routine: Pre-Market (08:30–09:15), Intraday Monitoring (09:15–15:30), Post-Close Analysis (15:45–17:00), and Weekend Review.

Why It Exists

Prevents reactive decision-making and ensures disciplined interaction with quantitative market intelligence.

When to Use:Follow daily during active trading weeks.
How to Use:Pre-market: check Discover & Overnight Filings. Intraday: monitor Alerts & Intraday Snapshots. Post-close: verify EOD publication & update watchlists.
Expected Outcome:Consistent, error-free daily market intelligence cycle.
Important Operational Invariants & Rules:
  • Autonomous EOD runs automatically post-close (16:00 IST) β€” manual pipeline execution is NOT required.
  • Always check Market Breadth in Insights before initiating new aggressive swing setups.
View-Only Scope:The daily schedule is automated; users observe rather than manually triggering EOD.
Safe to Change:Alert acknowledgment state, active watchlist composition.
Never Touch:Do not attempt manual session rollovers during active market hours.
βœ“ Healthy Normal State:EOD publication completed by 16:30 IST with status SUCCESS.
⚠️ When Something Is Wrong:Publication gate blocked past 17:00 IST requiring operator catchup.
Example: After the Control Plane publishes the completed EOD session, the analyst reviews the resulting Top 10 Swing candidates.
Next: Explore 'Discover' to examine precomputed candidate boards.Open Discover
#daily workflow#pre-market#intraday routine#post-close#schedule#routine#workflow steps

3. Discover

Intelligence & AnalyticsVIEW ONLY
What It Is

The flagship opportunity discovery surface displaying precomputed, model-ranked candidate boards tailored for 4 distinct time horizons: Swing, Positional, Long Term, and Intraday.

Why It Exists

Eliminates the need for manual parameter filtering by precomputing multi-factor technical setups across the entire 8,400+ equity universe.

When to Use:Use at the start of every trading session to discover top quantitative trade ideas.
How to Use:Select your desired investment horizon tab (e.g. Swing: 5–20 sessions). Review top-ranked cards, risk-reward ratios, and target levels.
Expected Outcome:Rapid identification of the highest-confluence setups in the market.
Important Operational Invariants & Rules:
  • Discover vs Scanner: Discover is curated and precomputed; Scanner allows custom multi-parameter screening.
  • Displayed Count (Top-N): Limits card display to the top 15 candidates; total candidates may be higher.
View-Only Scope:Candidate scoring and ranking are computed offline and strictly view-only.
Safe to Change:Horizon tab selection, adding candidates to watchlists.
Never Touch:Do not confuse target price with guaranteed exit orders.
βœ“ Healthy Normal State:Candidates populated with clear ranks, positive R:R ratios, and verified data provenance.
⚠️ When Something Is Wrong:Empty boards on a normal trading session indicates EOD scanner pipeline did not run.
Example: Discover Swing tab displays TCS at Rank #1 with Setup Score 82.4/100, R:R 2.4:1, and Target +8.5%.
Next: Open 'Scanner' if you wish to adjust indicator thresholds manually.Open Scanner
#discover#discover vs scanner#horizons#swing#positional#long term#opportunities#precomputed

4. Scanner

Intelligence & AnalyticsACTIONABLE
What It Is

Interactive quantitative screening engine allowing operators and analysts to customize strategy criteria, technical indicator thresholds, and sub-universe filters.

Why It Exists

Empowers deep quantitative research by exposing underlying strategy rules, moving average filters, and momentum parameters.

When to Use:Use when you have specific technical hypotheses (e.g. finding stocks with RSI 50–65 and 2.0x volume surge).
How to Use:Select setup board, choose quality tier, adjust min score or indicator sliders, and review the filtered results table.
Expected Outcome:Filtered table of matching securities with complete rule breakdown transparency.
Important Operational Invariants & Rules:
  • Evaluated Scope: Number of securities examined after universe eligibility gates.
  • Hard Filter Qualified: Number of securities satisfying all mandatory technical criteria.
  • Top-N Displayed: The subset of qualified candidates displayed on the page (default 20, 50, 100).
  • Top-N limits UI rendering only; it never truncates universe evaluation!
View-Only Scope:Scanning evaluates precomputed parquet datasets; no live remote queries are executed.
Safe to Change:Strategy sliders, category filters, exchange toggles, sort orders.
Never Touch:Do not attempt to execute trades directly from the scanner table.
βœ“ Healthy Normal State:Instant responsive filtering (< 100ms) with accurate candidate counts.
⚠️ When Something Is Wrong:Evaluated scope = 0 indicates universe metadata failure.
Example: User sets Min Score = 70, Category = EMA, Exchange = NSE; Scanner evaluates 7,550 stocks and returns 34 qualified.
Next: Click 'Detail' on any row to open Stock Intelligence.Open Stock Detail
#scanner#evaluated scope#qualified#top n#hard filters#screening#strategy criteria#scanner vs discover

5. Stock Intelligence

Intelligence & AnalyticsVIEW ONLY
What It Is

Deep-dive analytical intelligence dossier for an individual security, featuring multi-timeframe charts, technical indicators, pivot targets, invalidation stops, and corporate history.

Why It Exists

Provides complete 360-degree technical and contextual diligence before committing capital.

When to Use:Use whenever analyzing a specific stock identified in Discover or Scanner.
How to Use:Search symbol or click candidate row. Review Technical Setup, Multi-Timeframe Matrix, Classic Pivot Levels, and Risk Parameters.
Expected Outcome:Holistic understanding of price structure, support/resistance, and active alert events.
Important Operational Invariants & Rules:
  • All levels (Target, Stop, Pivot) are analytical references, NOT trade recommendations.
  • Dual-listed securities display unified company identity with NSE/BSE toggle.
View-Only Scope:Completely view-only analytical research dossier.
Safe to Change:Chart timeframe tabs (Daily, Weekly, Monthly), adding to Watchlist.
Never Touch:Never interpret target price as a guarantee of execution.
βœ“ Healthy Normal State:Complete OHLCV history, validated indicator values, and clear setup status.
⚠️ When Something Is Wrong:UNAVAILABLE banner appears when price data is missing or company is suspended.
Example: Viewing TCS: Close β‚Ή4,150, EMA 55 at β‚Ή4,020 (support), Classic R2 at β‚Ή4,480 (target), Setup Score 82.4.
Next: Add the security to Watchlists to track future developments.Open Watchlists
#stock intelligence#stock detail#charts#technical indicators#pivots#stop loss#target price

6. Watchlists

Intelligence & AnalyticsACTIONABLE
What It Is

User-curated portfolios and monitoring lists for organizing securities of interest.

Why It Exists

Allows focused observation of prioritized candidates without searching repeatedly.

When to Use:Use to group candidates by theme, sector, or setup readiness.
How to Use:Create named watchlists (e.g. 'Tech Breakouts'). Add stocks from Discover or Scanner. Monitor consolidated metrics.
Expected Outcome:Custom organized monitoring board with quick links to stock analytics.
Important Operational Invariants & Rules:
  • CRITICAL INVARIANT: Adding a stock to a watchlist does NOT automatically deliver Telegram alerts!
  • Telegram delivery requires configuring an alert rule or recipient group subscription.
View-Only Scope:Underlying market data in watchlist cards is read-only.
Safe to Change:Create, rename, add/remove symbols, delete watchlists.
Never Touch:Do not assume watchlists execute stop loss orders.
βœ“ Healthy Normal State:Watchlist loads instantly with current prices and setup badges.
⚠️ When Something Is Wrong:Removed stock continues appearing (clear local browser cache).
Example: User creates 'Banking Swing' with HDFCBANK, ICICIBANK, and SBIN; tracks combined average score.
Next: Open 'Alerts' to configure custom notification rules for your watchlist.Open Alerts Center
#watchlists#portfolios#monitoring#custom lists#symbols#tracking

7. Alerts

Alerting & NotificationsVIEW ONLY
What It Is

Centralized operational intelligence feed aggregating all quantitative signals, corporate filings, earnings announcements, corporate actions, and system health events.

Why It Exists

Consolidates diverse market events into a single prioritized, searchable stream.

When to Use:Check continuously during market hours and post-close for new intelligence items.
How to Use:Filter by Category tab (Signals, Filings, Actions, etc.), filter by Severity (Critical, High, Medium, Info), or search by symbol.
Expected Outcome:Comprehensive situational awareness of all significant market developments.
Important Operational Invariants & Rules:
  • Alert Generated != Telegram Sent: All alerts appear here in the web feed even when Telegram is in DRY_RUN!
  • Mark Read (Acknowledgment) archives items from unread counts without deleting history.
View-Only Scope:Alert records are immutable; acknowledging updates local read status only.
Safe to Change:Severity filters, category tabs, search queries, acknowledge button.
Never Touch:Do not delete backend alert logs directly from the filesystem.
βœ“ Healthy Normal State:Feed populated with real timestamped events; search returns instant matches.
⚠️ When Something Is Wrong:Empty feed on a market day indicates alert selection engine did not run.
Example: Alert feed displays HIGH severity 'QUARTERLY_RESULT: TCS' alongside MEDIUM 'BREAKOUT: RELIANCE'.
Next: Review 'Alert Types' to understand the exact meaning of every alert category.Open Alert Feed
#alerts#alert feed#how alerts work#notifications#unread#mark read#severity filter

8. Alert Types

Alerting & NotificationsVIEW ONLY
What It Is

Authoritative reference defining all 41 supported EGTS alert types across technical signals, financial results, corporate actions, and system health.

Why It Exists

Eliminates ambiguity so analysts know exactly what generated an event, why it matters, and what to do next.

When to Use:Consult whenever encountering an unfamiliar alert category in the feed or outbox.
How to Use:Search by category name or browse by domain tab in the Alert Types Modal.
Expected Outcome:Immediate clarity on generator origin, materiality, and recommended action.
Important Operational Invariants & Rules:
  • MATERIAL_CORPORATE_EVENT: High-materiality disclosures passing regulatory significance gates.
  • QUARTERLY_RESULT vs RESULTS_AVAILABLE: Result is the filing event; Available confirms normalized statements are parsed.
  • DAILY_RECENT_TOP: Flags top setups in newly listed equities / IPOs.
View-Only Scope:Alert category enums are strictly defined in backend models (`src/alerts/models.py`).
Safe to Change:Informational reference; completely safe to browse.
Never Touch:Never invent or synthesize fictional alert types.
βœ“ Healthy Normal State:Every generated alert in the feed maps to an authoritative definition.
⚠️ When Something Is Wrong:Unrecognized alert type appearing in feed (indicates undocumented backend enum).
Example: Looking up MATERIAL_CORPORATE_EVENT: Generator = Corporate Action Materiality Filter; Severity = HIGH; Action = Read disclosure.
Next: Learn how to build custom triggers in 'Custom Rules'.Open Alerts Center
#alert types#material corporate event#quarterly result#daily recent top#results available#enums#categories

9. Custom Rules

Alerting & NotificationsACTIONABLE
What It Is

Interactive rule builder allowing operators and users to create personalized alert triggers based on setup board, score, volume, and moving averages.

Why It Exists

Tailors intelligence delivery to individual trading styles and specific watchlist criteria.

When to Use:Use when you want automated notifications for specific technical conditions.
How to Use:Click 'New Rule'. Define name, target scope (All, Watchlist, Specific Symbols), min score, and indicator filters. Save rule.
Expected Outcome:Rule is registered in RuleStore and evaluated on every market data cycle.
Important Operational Invariants & Rules:
  • Rule evaluation is completely safe and operates in DRY_RUN baseline.
  • Destructive action: Deleting a rule opens a confirmation dialog detailing permanent removal.
View-Only Scope:Underlying indicator calculations evaluated by rules are read-only.
Safe to Change:Create, enable/disable, modify criteria, and delete custom rules.
Never Touch:Do not create 0-cooldown rules that trigger dozens of times per hour.
βœ“ Healthy Normal State:Rules show status ACTIVE with verified evaluation timestamps.
⚠️ When Something Is Wrong:Rule fails to trigger (verify whether criteria are too restrictive).
Example: Rule 'RSI Bullish Breakout': Score >= 70, RSI > 55, RVOL > 2.0x, Scope = Watchlist 'Tech'.
Next: Configure 'Telegram Recipients' to receive your custom alerts externally.Open Rule Builder
#custom rules#rule builder#create alert#filter group#thresholds#custom alerts

10. Telegram Recipients

Alerting & NotificationsACTIONABLE
What It Is

Recipient management console for binding Telegram accounts, managing group memberships (TRADER, ADMIN, OPERATIONS), and setting quiet hours.

Why It Exists

Ensures alerts reach the correct team members with appropriate severity filtering and zero notification spam.

When to Use:Use when onboarding a new team member or updating notification preferences.
How to Use:Add recipient, generate token, send `/start <token>` to the EGTS bot on Telegram, and verify status flips to VERIFIED.
Expected Outcome:Verified recipient with bound Telegram Chat ID ready for delivery.
Important Operational Invariants & Rules:
  • Zero Secret Leaks: Bot tokens and full chat IDs are never serialized to the client browser.
  • Admin Mirror: Operators in group ADMIN automatically receive mirrored critical system alerts.
  • Destructive action: Deleting a recipient prompts for confirmation stating delivery will cease.
View-Only Scope:Bot credentials are read from server environment variables and cannot be altered from UI.
Safe to Change:Add recipient, edit minimum severity, toggle enabled/disabled, generate binding token.
Never Touch:Do not share verification tokens publicly.
βœ“ Healthy Normal State:Recipients show VERIFIED with masked chat IDs (`1234***`).
⚠️ When Something Is Wrong:Recipient stuck in PENDING (user has not clicked /start on the bot).
Example: Operator adds recipient 'Ronak (Admin)', binds via Telegram /start; status becomes VERIFIED.
Next: Review 'Telegram Modes' to understand delivery gating.Open Recipient Manager
#telegram recipients#how to add telegram user#recipient binding#verification token#admin mirror#chat id

11. Telegram Modes

Alerting & NotificationsVIEW ONLY
What It Is

Hierarchical 4-tier production delivery safety gate: DRY_RUN, TEST_RECIPIENT (Canary only), LIMITED_LIVE (Authoritative Dispatch to verified Admin/Operators), and LIVE (Full broadcast).

Why It Exists

Prevents accidental notification blasts to traders or clients during testing, staging, and maintenance.

When to Use:Check current mode before initiating test evaluations or commissioning.
How to Use:View current mode in Delivery Audit card. Production authoritative dispatch operates in LIMITED_LIVE (with legacy CP notification adapter reporting DRY_RUN).
Expected Outcome:Complete assurance that operational and qualified events are delivered strictly to verified Admin/Operators.
Important Operational Invariants & Rules:
  • DRY_RUN: Simulation mode. Evaluates rules, populates Outbox simulation, logs audit trail, but makes ZERO external network requests.
  • TEST_RECIPIENT: Dispatches ONLY to recipients with is_test_recipient = True (Canary).
  • LIMITED_LIVE (Authoritative): Dispatches critical operational alerts and qualified events strictly to verified Admin/Operator recipients.
  • LIVE: Full multi-recipient broadcast mode.
  • Legacy CP Adapter: Legacy control-plane notification adapter may still report DRY_RUN; do not confuse it with authoritative LIMITED_LIVE alert dispatch.
View-Only Scope:Delivery mode is governed by `TELEGRAM_DELIVERY_MODE` environment variable and is view-only in UI.
Safe to Change:Trigger safe Dry-Run Evaluation button in Outbox view.
Never Touch:Never promote mode to LIVE without verified Canary recipient and delivery audit proof.
βœ“ Healthy Normal State:AUTHORITATIVE TELEGRAM: LIMITED_LIVE (VERIFIED ADMIN) displayed with green check badge; Legacy CP Adapter: DRY_RUN.
⚠️ When Something Is Wrong:Mode changed to LIVE without operator commissioning.
Example: In LIMITED_LIVE mode, operational alerts (such as DAILY_PUBLICATION_SUCCESS) and qualified setups are dispatched securely to verified Admin Telegram.
Next: Inspect 'Delivery Outbox & Audit' to see how messages are queued and logged.Open Delivery Audit
#telegram modes#dry run#test recipient#limited live#live#safety gates#canary#what does limited live mean#legacy cp adapter

12. Delivery Outbox & Audit

Alerting & NotificationsVIEW ONLY
What It Is

Transactional FIFO delivery queue and historical audit ledger tracking every generated message, matched recipient, suppression reason, and dispatch attempt.

Why It Exists

Provides complete forensic auditability for every notification generated by the platform.

When to Use:Use to verify why a recipient received (or did not receive) an alert, or to inspect queue status.
How to Use:Open Outbox tab. Inspect 'Pending in Outbox', 'Recent Delivery Items', and 'Routing & Suppression Audit History'.
Expected Outcome:100% transparent audit lineage for every alert event.
Important Operational Invariants & Rules:
  • Pending Outbox in DRY_RUN: Items are staged for simulation. No external Telegram delivery occurs while mode is DRY_RUN.
  • Suppression Reasons: Documented transparently (e.g. 'Below minimum severity', 'Quiet hours', 'Cooldown active').
  • SAFETY DEFECT NOTE: Old pending items from dry runs must be cleared before mode promotion to prevent sudden dispatch.
View-Only Scope:Audit trail is an append-only JSONL log (`data/state/alerts/delivery_audit.jsonl`).
Safe to Change:Trigger safe 'Run Dry-Run Evaluation' or 'Refresh Audit Data'.
Never Touch:Do not manually edit delivery_outbox.jsonl.
βœ“ Healthy Normal State:Pending count is low or reflects recent dry-run evaluation; audit records show clear reasons.
⚠️ When Something Is Wrong:High queue backlog with status FAILED (indicates network or bot token misconfiguration).
Example: Delivery Audit shows Alert ALT_FND_TCS matched Recipient Operator (SENT: DRY_RUN) and suppressed for Trader (Quiet Hours).
Next: Explore 'Intelligence Digests' for consolidated scheduled summaries.Open Delivery Audit View
#delivery outbox#pending outbox#audit records#what does pending outbox mean#suppression#cooldown#quiet hours#deduplicated

13. Intelligence Digests

Alerting & NotificationsVIEW ONLY
What It Is

Scheduled consolidated briefing engine grouping multiple domain events into one structured report (Morning Brief, Hourly Market Movers, Daily EOD Summary).

Why It Exists

Reduces notification fatigue by bundling hundreds of individual alerts into periodic high-level intelligence summaries.

When to Use:Read at 08:30 IST (Morning Brief) and 17:00 IST (Daily EOD Summary).
How to Use:Select digest format (Morning Brief, EOD Summary) from dropdown; review domain counts and structured markdown text.
Expected Outcome:Consolidated, executive overview of market intelligence across all sectors.
Important Operational Invariants & Rules:
  • Digest vs Alert Feed: Feed is real-time and granular; Digest is scheduled and consolidated.
  • Freshness Transparency: If the latest generated digest session is older than platform EOD, the UI explicitly displays a STALE warning badge.
View-Only Scope:Digests are synthesized from historical alert feed store records and are view-only.
Safe to Change:Digest format dropdown selection, refresh button.
Never Touch:Do not hardcode localhost or production IP addresses in digest footers.
βœ“ Healthy Normal State:Digest session matches current platform EOD session.
⚠️ When Something Is Wrong:Digest displays STALE warning when digest generation has not run since earlier session.
Example: Morning Intelligence Brief summarizes 42 items: 12 Signals, 8 Filings, 14 Financial Results, 8 System alerts.
Next: Visit 'Insights' to inspect market-wide breadth and sector trends.Open Digest Preview
#digest#intelligence digest#morning brief#eod summary#old digest#stale digest#digest freshness

14. Insights

Intelligence & AnalyticsVIEW ONLY
What It Is

Macro and market-wide intelligence cockpit featuring 7 specialized analytical panels: Market Regime, Sector Relative Strength, Advance-Decline Breadth, Volatility Regimes, Fundamentals Availability, Historical Simulation, and System Anomalies.

Why It Exists

Prevents isolated stock picking by contextualizing setups within macro market breadth and sector flows.

When to Use:Use before taking new positions to ensure market breadth supports long exposure.
How to Use:Review Tab 1 (Market Breadth) to check net advancing stocks. Check Tab 2 (Sectors) for leading sectors. Review Tab 5 (Fundamentals).
Expected Outcome:Clear macro perspective on whether current market environment is risk-on, neutral, or risk-off.
Important Operational Invariants & Rules:
  • System Anomalies panel reports operational data and runtime anomalies β€” it is NOT a trading recommendation!
  • Model Score is a ranking metric, NOT a calibrated win probability.
View-Only Scope:All breadth indices and sector relative strength calculations are read-only.
Safe to Change:Analytical tab navigation.
Never Touch:Do not trade contrary to severe risk-off breadth regimes without hedging.
βœ“ Healthy Normal State:All tabs load with current session timestamps and consistent, explicitly labelled listing/company populations.
⚠️ When Something Is Wrong:Breadth metrics show UNAVAILABLE if market data parquet files are missing.
Example: Market Breadth shows 68% advancing issues with Auto and IT leading sector relative strength.
Next: Read 'Fundamentals / Context Availability' to understand data coverage.Open Insights Page
#insights#market breadth#advance decline#sector strength#macro regime#anomalies

15. Fundamentals / Context Availability

Data Architecture & GovernanceVIEW ONLY
What It Is

Authoritative architectural explanation of the three distinct fundamentals and disclosure layers in EGTS.

Why It Exists

Clarifies why 'Source Registry Not Configured' does NOT mean zero usable fundamentals exist in EGTS.

When to Use:Consult when reviewing company earnings, financial statements, or exchange filings.
How to Use:Inspect the 3 distinct tiers: Live Provider Registry, Materialized Offline Fundamentals, and Contextual Alert Stream.
Expected Outcome:Accurate understanding of data provenance and the current company-level coverage reported by the API.
Important Operational Invariants & Rules:
  • Layer A (Live Streaming Provider Registry): NOT COMMISSIONED / SOURCE LIMITED (0/4 active streaming feeds in production baseline).
  • Layer B (Materialized Offline Fundamentals): PARTIALLY AVAILABLE (Normalized financial statements, shareholdings, and ratios are preserved offline for covered equities like TCS).
  • Layer C (Corporate Disclosure Stream): AVAILABLE (Filings, results, and corporate actions processed during scan passes).
View-Only Scope:Source registry commissioning is an infrastructure configuration and view-only.
Safe to Change:Completely safe reference view.
Never Touch:Do not run unverified scraping scripts against exchange servers.
βœ“ Healthy Normal State:Truthful reporting of uncommissioned live feeds alongside available offline normalized files.
⚠️ When Something Is Wrong:Claiming live feeds are active when no streaming daemon is running.
Example: TCS displays valid normalized financial statements (Layer B) even though the live external streaming feed (Layer A) is uncommissioned.
Next: Inspect 'System Overview' for the broader platform architecture.Open Insights (Fundamentals Tab)
#fundamentals not configured#source limited#fundamentals#context availability#financial statements#xbrl#shareholding

16. System Overview

System Operations & Control PlaneVIEW ONLY
What It Is

Executive operations cockpit displaying overall platform health, release SHA identity, session alignment, and subsystem statuses.

Why It Exists

Provides single-pane-of-glass operational visibility for operators and system administrators.

When to Use:Check at system startup, deployment verification, and daily operational health checks.
How to Use:Review Git Commit SHA, Release Version, Platform EOD Session, and Control Plane Heartbeat in the top banner.
Expected Outcome:Immediate confirmation of platform operational integrity.
Important Operational Invariants & Rules:
  • Frontend and Backend SHA must match for release consistency.
  • Platform EOD session defines the canonical mathematical baseline for all scanner calculations.
View-Only Scope:Platform health metrics and version identifiers are strictly read-only.
Safe to Change:Tab switching between Overview, Operations, and Lineage.
Never Touch:Do not alter git tags or release commits directly on production servers.
βœ“ Healthy Normal State:Release SHA matches deployed tag; EOD session matches current date.
⚠️ When Something Is Wrong:SHA mismatch between frontend and backend indicates partial deployment.
Example: System Overview displays the runtime release, commit, current EOD session, and reported control-plane state.
Next: Inspect 'Architecture' for data flow pipelines.Open System Cockpit
#system overview#cockpit#system health#release version#git commit#platform status

17. Architecture

System Operations & Control PlaneVIEW ONLY
What It Is

Architectural blueprint detailing the end-to-end data pipeline from Raw Exchange Feeds βž” Silver Parquet βž” Gold Analytics βž” PublicationCURRENT βž” FastAPI βž” Next.js Frontend.

Why It Exists

Documents data immutability, directory isolation, and single-writer boundaries.

When to Use:Review when debugging data lineage or understanding cache invalidation.
How to Use:Trace the pipeline stages in the architecture diagram; inspect contract schemas.
Expected Outcome:Full comprehension of how raw ticks become published candidate boards.
Important Operational Invariants & Rules:
  • Immutability: Parquet partitions are write-once per session.
  • Atomic Publication: Data is promoted to `data/publication/CURRENT` atomically via manifest symlinks.
View-Only Scope:Architectural structure is enforced by code contracts.
Safe to Change:Informational documentation.
Never Touch:Never bypass publication gates to write directly into gold analytics.
βœ“ Healthy Normal State:Clean linear data flow with zero circular dependencies.
⚠️ When Something Is Wrong:Dangling symlinks in publication directory.
Example: EOD pipeline ingests raw bhavcopy, builds Silver OHLCV, computes Gold indicators, and publishes CURRENT manifest.
Next: Examine 'Dependencies / Lineage' for specific artifact dependencies.Open System (Architecture)
#architecture#data flow#pipeline stages#silver gold#publication current#fastapi#nextjs

18. Dependencies / Lineage

Data Architecture & GovernanceVIEW ONLY
What It Is

Granular dependency graph and artifact lineage tracker showing which raw datasets produce which analytical outputs.

Why It Exists

Ensures that if an upstream data defect occurs, operators can trace every affected downstream metric.

When to Use:Use during data reconciliation, restatement audits, or bug investigations.
How to Use:Search artifact by name to view upstream sources and downstream consumers.
Expected Outcome:Complete traceability from raw bhavcopy to front-end setup scores.
Important Operational Invariants & Rules:
  • Every candidate board traces back to specific parquet partitions and publication fingerprints.
View-Only Scope:Lineage maps are generated from code contracts.
Safe to Change:Browse lineage relationships.
Never Touch:Do not manually break file dependencies.
βœ“ Healthy Normal State:Zero unverified or untracked artifacts in publication path.
⚠️ When Something Is Wrong:Missing upstream dependency prevents downstream pipeline execution.
Example: A Discover candidate traces to the publication-session indicators and its company intelligence artifact.
Next: Review 'Listings' for security identifier mapping.Open System (Lineage)
#dependencies#lineage#data provenance#artifacts#upstream downstream#traceability

19. Listings

Data Architecture & GovernanceVIEW ONLY
What It Is

Unified master equity taxonomy managing current and historical NSE/BSE listings through ISIN and company identity frameworks.

Why It Exists

Prevents duplicate company accounting and handles dual-listed stocks without conflating security and company counts.

When to Use:Use when cross-referencing BSE security codes (e.g. 500325) with NSE symbols (RELIANCE).
How to Use:Search by Symbol, BSE Code, Company Name, or ISIN.
Expected Outcome:Authoritative listing metadata, tick size, lot size, and exchange classification.
Important Operational Invariants & Rules:
  • Shared Company Entity: TCS on NSE and TCS on BSE map to canonical company entity `comp-tcs`.
  • Dual-listed stocks share fundamental filings while maintaining independent exchange order books.
View-Only Scope:Listing taxonomy is derived from official exchange master directories.
Safe to Change:Exchange view toggles.
Never Touch:Never invent or alter canonical ISIN mappings.
βœ“ Healthy Normal State:Runtime current-listing and unique-company counts reconcile with the identity registry.
⚠️ When Something Is Wrong:Unmapped listing resulting in duplicate or fragmented analytics.
Example: Searching '500325' resolves to RELIANCE on BSE, linked to RELIANCE on NSE via ISIN INE002A01018.
Next: Learn about operational automation in 'Operations & Control Plane'.Open Scanner
#listings#isin#dual listed#nse bse mapping#symbol lookup#company identity

20. Operations & Control Plane

System Operations & Control PlaneDIAGNOSTIC ONLY
What It Is

The single-writer autonomous operations daemon (`egts-control-plane`) responsible for scheduled pipeline execution, health monitoring, and session management.

Why It Exists

Guarantees race-free data processing by enforcing single-writer locks across all data partitions.

When to Use:Check daemon heartbeat, active job status, and lock status in the System Cockpit.
How to Use:Monitor the 'Scheduler Authority' card. Verify: CONTROL PLANE: ACTIVE (GOOD) and LEGACY SCHEDULER: MASKED/OFF (GOOD).
Expected Outcome:Uninterrupted autonomous operations with fresh heartbeats.
Important Operational Invariants & Rules:
  • LEGACY SCHEDULER MASKED/OFF = GOOD: The legacy egts-scheduler systemd service is intentionally disabled to avoid dual-writer race conditions!
  • NEVER instruct anyone to start egts-scheduler!
  • Control Plane owns all internal scheduling autonomously.
View-Only Scope:Control Plane daemon status is monitored via IPC heartbeat files.
Safe to Change:Diagnostic viewing.
Never Touch:Do not unmask or start the legacy egts-scheduler service.
βœ“ Healthy Normal State:CONTROL PLANE: ACTIVE (Heartbeat < 60s); LEGACY SCHEDULER: MASKED/OFF.
⚠️ When Something Is Wrong:CONTROL PLANE: INACTIVE or STALE HEARTBEAT (restart daemon service via systemctl).
Example: Control Plane heartbeat updated 12 seconds ago; legacy scheduler confirmed masked in systemd.
Next: Check 'System Alerts' for automated operational warnings.Open System Cockpit
#control plane#scheduler masked#legacy scheduler#single writer#heartbeat#why scheduler masked#daemon

21. System Alerts

System Operations & Control PlaneVIEW ONLY
What It Is

Automated operational health and exception alerts generated by the platform watchdog (Pipeline failures, feed degradations, disk warnings, stale data).

Why It Exists

Ensures operators are immediately alerted to infrastructure or data quality anomalies before users notice.

When to Use:Review during daily operations check or when debugging pipeline delays.
How to Use:Filter Alert Feed by category tab 'System' or check the System Cockpit alerts drawer.
Expected Outcome:Immediate identification of system bottlenecks or failing pipeline stages.
Important Operational Invariants & Rules:
  • System alerts are automatically mirrored to Admin Telegram recipients when configured.
  • CRITICAL alerts indicate that operator intervention is required.
View-Only Scope:System alerts are generated by internal health sentinels and are read-only.
Safe to Change:Acknowledge resolved system alerts.
Never Touch:Do not suppress CRITICAL system alerts without investigating root cause.
βœ“ Healthy Normal State:Zero unacknowledged CRITICAL system alerts.
⚠️ When Something Is Wrong:Active OPERATOR_PIPELINE_FAILURE or EOD_FAILED alerts requiring runbook action.
Example: OPERATOR_PIPELINE_FAILURE alert fired with exit code 1; operator opens runbook to diagnose.
Next: Consult 'Runbook' for exact resolution commands.Open Alert Feed (System Tab)
#system alerts#pipeline failure#provider down#eod failed#data stale warning#operational alerts

22. Runbook

System Operations & Control PlaneADVANCED OPERATOR ACTION
What It Is

Authoritative operator runbook containing standard operational procedures, diagnostic terminal commands, backup/restore sequences, and recovery actions.

Why It Exists

Ensures consistent, safe operations without improvising terminal commands during incidents.

When to Use:Use when diagnosing failed jobs, creating backups, or executing planned maintenance.
How to Use:Locate the specific incident scenario; copy and run the verified operator commands in the terminal.
Expected Outcome:Fast, deterministic incident resolution without accidental data loss.
Important Operational Invariants & Rules:
  • Commands must be executed from the workspace root.
  • Never run modifying commands without verifying current READ_ONLY and single-writer locks.
View-Only Scope:The runbook in UI is documentation; actions are executed by authorized operators in terminal.
Safe to Change:Runbook documentation reference.
Never Touch:Never execute destructive commands (e.g. `rm -rf data`) without an verified offline backup.
βœ“ Healthy Normal State:All automated daily checks pass without requiring manual runbook intervention.
⚠️ When Something Is Wrong:Failed pipeline requiring manual catchup script execution.
Example: To check Control Plane status: `launchctl list | grep egts`.
Next: Understand data timing in 'Data Freshness'.Open System (Operations Tab)
#runbook#operator commands#backup#restore#how do i backup#maintenance#troubleshooting commands

23. Data Freshness

Data Architecture & GovernanceVIEW ONLY
What It Is

Data freshness governance framework defining update SLAs, session schedules, and staleness detection across all data tiers.

Why It Exists

Guarantees that users never make decisions on outdated or misleading historical prices.

When to Use:Consult when checking whether today's closing prices have been incorporated.
How to Use:Check 'Platform EOD Session' and 'Freshness' badges displayed in the top header and on cards.
Expected Outcome:Transparent, truthful indication of data timeliness.
Important Operational Invariants & Rules:
  • Platform EOD Session: The latest official exchange session incorporated into the analytical model.
  • Closed Market: Over weekends and exchange holidays, data remains at the latest traded session β€” this is NORMAL and not a bug!
View-Only Scope:Freshness evaluations compare file modification dates with the exchange trading calendar.
Safe to Change:Informational view.
Never Touch:Never spoof dates to make old data appear fresh.
βœ“ Healthy Normal State:Data Freshness badge displays 'CURRENT' during trading weeks.
⚠️ When Something Is Wrong:Data Freshness badge displays 'STALE' on a regular market evening.
Example: On a weekend, Platform EOD Session displays the latest completed trading session with a market-closed label.
Next: Learn how session rollovers work in 'Sessions / CURRENT / LKG'.Open System Cockpit
#data freshness#freshness#stale#session date#market closed#sla#timeliness

24. Sessions / CURRENT / LKG

Data Architecture & GovernanceVIEW ONLY
What It Is

Atomic publication and rollback architecture managing transitions between session publications, CURRENT pointers, and Last Known Good (LKG) states.

Why It Exists

Ensures that a failed EOD pipeline never leaves the platform in a corrupted or half-baked state.

When to Use:Consult during EOD publication verification or when investigating rollbacks.
How to Use:Inspect `data/publication/current_publication_manifest.json`. If a gate fails, the system safely retains the LKG publication.
Expected Outcome:Guaranteed zero downtime and zero partial-data exposure to front-end users.
Important Operational Invariants & Rules:
  • CURRENT: The active verified publication manifest symlink serving production API requests.
  • LKG (Last Known Good): The previous fully validated session preserved as a fallback rollback target.
  • Publication Gate: Enforces 8 mathematical and taxonomy invariants before promoting new data to CURRENT.
View-Only Scope:Publication pointer promotion is managed autonomously by the Control Plane.
Safe to Change:View manifest metrics.
Never Touch:Do not manually redirect the CURRENT symlink to an unverified session folder.
βœ“ Healthy Normal State:Checks Passed: 8/8 invariants valid; status SUCCESS.
⚠️ When Something Is Wrong:Checks Passed < 8; publication held in quarantine while LKG continues serving.
Example: Current manifest shows checks passed: CANONICAL_SESSION_MATCH, MATHEMATICAL_INVARIANTS_VALID, REBASE_TAXONOMY_CLEAR.
Next: Browse 'Common Terms' for full glossary definitions.Open System Cockpit
#what is lkg#lkg#current#publication session#publication manifest#rollback#atomic publication

25. Common Terms

Getting Started & WorkflowVIEW ONLY
What It Is

Authoritative terminology glossary containing 42+ standardized definitions across quantitative modeling, data governance, operations, and Telegram routing.

Why It Exists

Establishes a single unambiguous vocabulary for all engineers, operators, and quantitative analysts.

When to Use:Consult whenever encountering technical terms like RVOL, EMA 55, Brier Score, Cooldown, or Dual-Listed.
How to Use:Browse by category (Model, Data, Operations, Telegram) or use the search bar to locate specific terms.
Expected Outcome:Immediate conceptual clarity with backend semantics and healthy/warning state examples.
Important Operational Invariants & Rules:
  • Every term includes formal backend semantics explaining exactly how it is computed in the source code.
View-Only Scope:Glossary definitions are maintained in `apps/web/lib/knowledge.ts`.
Safe to Change:Browse and search terms.
Never Touch:Do not redefine established statistical terms arbitrarily.
βœ“ Healthy Normal State:All glossary terms cross-referenced with working in-product links.
⚠️ When Something Is Wrong:Missing term in search index.
Example: Brier Score: Mean squared error between probabilistic forecast and binary outcome; 0.0 is perfect.
Next: Review 'Normal User vs Operator Actions' for operational role boundaries.Open Guide (Glossary Tab)
#common terms#glossary#definitions#terminology#dictionary#vocabulary

Interactive Glossary Explorer

UniverseUNIVERSE

The total canonical collection of all known listed equity securities across NSE and BSE.

Authoritative immutable parquet store tracking current and historical securities with distinct status populations.
Source PresentDATA

Listings that have verifiable source records present in active exchange master files or historical archives.

Securities having raw Bhavcopy archives or exchange master records present on disk.
Current ListingUNIVERSE

Active equity securities currently traded on the exchange with valid trading status.

Listings with active trading status, excluding suspended, delisted, or uncommissioned series.
EligibleUNIVERSE

Securities meeting baseline operational and exchange requirements.

Active equity series (EQ, BE, A, B) with valid ISINs and sufficient historical depth.
Evaluated ScopeSCANNER

The deterministic subset of securities selected for evaluation under active filters.

Denominator of securities tested after applying exchange, board, instrument, and liquidity constraints.
Data ReadyDATA

Securities possessing authentic historical and current OHLCV bars and required indicator arrays.

Securities with complete, continuous indicator cache on the requested timeframe.
Hard Filter QualifiedSCANNER

Candidates that strictly satisfy 100% of the active hard criteria.

Deterministic count of securities passing every active AST condition.
DisplayedSCANNER

The final list of candidate rows rendered in the UI, bounded by min(Qualified, Top-N).

Presentation limit only; does not affect or reduce underlying evaluation.
Top-NSCANNER

A presentation-layer display limiter that caps the number of top-ranked rows rendered.

Slices ranked qualified results for UI rendering; never restricts universe evaluation.
SetupMODEL

A recognized quantitative technical pattern satisfying deterministic mathematical rules.

Named pattern (e.g. EMA 55/144 Pullback, RSI > 50 Momentum) evaluated against OHLCV.
Composite ScoreMODEL

A weighted multi-factor ranking index (0–100) combining trend, momentum, and volume.

Heuristic ranking metric used to order qualified candidates in read models.
Score != probabilityMODEL

Invariant stating that a composite model score is a ranking index, not an empirical probability.

Uncalibrated heuristic setup scores are strictly forbidden from being displayed as percentage win odds.
TargetRISK

A quantitative resistance level (Classic Pivot R2/R3, Swing High, or ATR projection).

Analytical resistance price level used to measure upside distance.
StopRISK

An objective structural invalidation level where the technical setup is invalidated.

Analytical invalidation price (swing low, EMA 55 baseline, or ATR buffer).
Risk:RewardRISK

The mathematical ratio of upside potential to structural target vs downside to stop loss.

Computed as (Target - Entry) / (Entry - Stop).
TimeframeSCANNER

The discrete bar aggregation interval used for calculation.

Supported: 5m, 15m, 1h, 4h (intraday); daily, weekly, monthly (macro).
MTFSCANNER

Multi-Timeframe analysis evaluating confluence across multiple bar regimes simultaneously.

Joint condition evaluation across higher, intermediate, and lower timeframe arrays.
PIT / Point-in-TimeDATA

Strict temporal isolation preventing lookahead bias by evaluating models as of snapshot dates.

Calculations only ingest data timestamps strictly <= snapshot session timestamp.
CURRENTDATA

The authoritative active publication pointer designating the latest materialized session.

Symlink or atomic pointer pointing to the most recently completed EOD read model directory.
LKGGOVERNANCE

Last Known Good publication artifact preserved as a fail-safe fallback.

Immutable backup snapshot retained to serve reads if current generation fails verification.
Publication SessionDATA

The specific calendar trade date represented by the currently published read model.

Date string (YYYY-MM-DD) embedded in read model metadata.
Latest Traded SessionDATA

The calendar trade date of the most recent market trading session completed on the exchange.

Determined from official NSE/BSE trading calendars.
Publication FingerprintGOVERNANCE

An immutable cryptographic hash uniquely identifying dataset lineage and parameters.

SHA-256 hash computed over publication artifacts and code version.
Read ModelDATA

A precomputed, denormalized read-optimized data artifact designed for rapid querying.

JSON/Parquet file written to data/published/ and served directly by endpoints.
MaterializedDATA

Data artifacts precomputed and written to disk by the pipeline rather than generated on the fly.

Persisted files on disk ready for static reads.
Source LimitedDATA

Status indicating external data sources did not provide complete records for a specific feed.

Handled gracefully with transparent limitation flags rather than fabrication.
Not ConfiguredOPERATIONS

Component state indicating optional integration parameters are intentionally omitted.

Readiness checks return NOT_CONFIGURED without throwing fatal exceptions.
StaleDATA

Data or cache timestamp older than the expected update schedule or publication threshold.

Evaluated by comparing artifact timestamp with market close schedule.
UnavailableDATA

Truthful indicator that data does not exist or could not be loaded; never synthesized.

API returns structured UNAVAILABLE response with explicit reason.
Control PlaneOPERATIONS

The authoritative autonomous automation daemon (egts-control-plane) for all pipelines.

Sole writer authority enforcing single-writer locks, session state machine, and scheduling.
Legacy SchedulerOPERATIONS

Legacy egts-scheduler systemd service intentionally INACTIVE and MASKED.

Masked to prevent dual-writer conflicts with the autonomous Control Plane.
DRY_RUNTELEGRAM

Safe notification dispatch mode evaluating rules and outbox audits without external sends.

Evaluates rules, logs outbox records, but makes zero network requests to Telegram.
TEST_RECIPIENTTELEGRAM

Controlled testing mode dispatching notifications exclusively to verified Canary/Admin recipients.

Filters outbox sends strictly to recipients with is_test_recipient == True.
LIMITED_LIVETELEGRAM

Gated live delivery mode restricting sends to specific critical channels or verified groups.

Restricts delivery to high-severity groups and verified members.
LIVETELEGRAM

Fully operational live delivery mode sending eligible alerts to all active verified recipients.

Dispatches to all matching recipients with verification_status == VERIFIED.
Canary RecipientTELEGRAM

A verified administrator designated to receive staging or pre-release notifications.

Recipient with is_test_recipient=True and groups=['ADMIN'].
Admin MirrorTELEGRAM

Safety routing rule where critical system notifications are mirrored to administrative recipients.

System operator alerts automatically append GRP_ADMIN to target recipient list.
Quiet HoursTELEGRAM

Time window (default 22:00–07:00 IST) suppressing non-critical notifications.

Alerts below CRITICAL severity are suppressed during configured hours.
CooldownTELEGRAM

The minimum mandatory quiet interval between repeated alerts for the same symbol and rule.

Checked against alert history store; subsequent alerts within window are suppressed.
OutboxTELEGRAM

The transactional queue where generated alerts are recorded before safe dispatch.

JSONL queue storing DeliveryOutboxItem instances with retry and attempt counters.
SuppressedTELEGRAM

Status of an alert withheld from external dispatch due to quiet hours, cooldown, or preferences.

Recorded in DeliveryAuditRecord with explicit suppression reason.
DeduplicatedTELEGRAM

Filter action that discards redundant identical alert events generated within an active window.

Identified via (symbol, rule_id, timeframe) fingerprint match.

26. Normal User vs Operator Actions

Safety & GovernanceVIEW ONLY
What It Is

Clear boundary separation distinguishing actions safe for normal market analysts versus privileged actions reserved exclusively for system operators.

Why It Exists

Prevents unauthorized or accidental execution of maintenance, backup, or service modification commands.

When to Use:Review to understand your permission boundaries within the platform.
How to Use:Normal users stay within Discover, Scanner, Stock Detail, Watchlists, and Alert Feed. Operators handle Control Plane, terminal runbooks, and backups.
Expected Outcome:Strict operational security and error-free multi-user governance.
Important Operational Invariants & Rules:
  • Normal User: Browse opportunities, filter candidates, manage personal watchlists, configure alert rules, view stock intelligence.
  • Operator: Monitor Control Plane daemon, execute backup/restore verification, audit outbox queues, manage Telegram delivery modes.
View-Only Scope:All operator commands require local terminal access and are barred from normal web UI.
Safe to Change:All Normal User capabilities.
Never Touch:Normal users must never run operational terminal scripts or modify daemon settings.
βœ“ Healthy Normal State:Clear UI badges (VIEW ONLY vs ACTIONABLE vs ADVANCED OPERATOR ACTION) throughout all pages.
⚠️ When Something Is Wrong:Unprivileged user attempting to execute operational maintenance.
Example: Adding a stock to a watchlist is a Normal User action; purging outbox queues is an Advanced Operator Action.
Next: Check 'Troubleshooting' for common issue resolutions.Open System Cockpit
#user vs operator#normal user#operator actions#roles#permissions#boundaries#safety levels

27. Troubleshooting

System Operations & Control PlaneDIAGNOSTIC ONLY
What It Is

Comprehensive diagnostic troubleshooting matrix covering common operational symptoms, root causes, and verified resolution steps.

Why It Exists

Enables rapid self-service problem resolution without escalation to engineering.

When to Use:Use whenever encountering unexpected UI states, blank fields, or missing alert deliveries.
How to Use:Find your symptom (e.g. 'Why is scheduler off?', 'Why is Telegram not sending?'). Follow the verified resolution checklist.
Expected Outcome:Swift return to normal healthy operating state.
Important Operational Invariants & Rules:
  • Scheduler Masked is NOT an error: Legacy scheduler must remain masked!
  • Telegram not sending: In DRY_RUN mode, external sending is intentionally suppressed by design.
  • Digest session older than EOD: Normal when new digest has not been generated post-close.
View-Only Scope:Troubleshooting guidance is diagnostic.
Safe to Change:Follow recommended resolution steps.
Never Touch:Do not execute force-kill commands without verifying lock ownership.
βœ“ Healthy Normal State:All symptoms resolve to known, documented healthy patterns.
⚠️ When Something Is Wrong:Unprecedented error requiring codebase inspection.
Example: Symptom: 'Why is Telegram not sending?' βž” Cause: Global mode is DRY_RUN βž” Resolution: Expected behavior in safe baseline.
Next: Understand safety constraints in 'Production Safety / READ_ONLY'.Open System Cockpit
#troubleshooting#why scheduler off#why telegram not sending#why digest old#common problems#faq#fixes

28. Production Safety / READ_ONLY

Safety & GovernanceVIEW ONLY
What It Is

Production safety architecture enforcing strict READ_ONLY operational mode, zero external network calls, zero broker connectivity, and credential protection.

Why It Exists

Guarantees that EGTS cannot execute unwanted financial orders, leak sensitive credentials, or mutate canonical market datasets.

When to Use:Review during compliance audits and security assessments.
How to Use:Verify READ_ONLY active badge in System Cockpit. Confirm Broker Execution = DISABLED in settings.
Expected Outcome:100% compliant, air-gapped analytical research environment.
Important Operational Invariants & Rules:
  • READ_ONLY: Canonical market data and publication partitions are strictly read-only for web processes.
  • Broker Execution DISABLED: There are ZERO broker APIs, order placement endpoints, or financial execution hooks.
  • Zero Secret Leaks: Bot tokens, chat IDs, and private keys are never exposed in API payloads or client code.
View-Only Scope:Safety mode is hardcoded into application contracts and cannot be overridden by UI flags.
Safe to Change:Compliance viewing.
Never Touch:Never attempt to attach live brokerage execution libraries to this codebase.
βœ“ Healthy Normal State:READ_ONLY: ENABLED; BROKER EXECUTION: DISABLED; ZERO SECRET LEAKS.
⚠️ When Something Is Wrong:Attempted write operation on canonical parquet directory.
Example: The platform confirms READ_ONLY mode active; attempting to place an order produces an architectural impossibility.
Next: Inspect remaining steps in 'Production Closeout / Remaining Acceptance'.Open System Cockpit
#read only#broker#production safety#zero secret leaks#air gapped#compliance#order execution disabled

29. Production Closeout / Remaining Acceptance

Safety & GovernanceDIAGNOSTIC ONLY
What It Is

Truthful operational status ledger detailing what is complete in this release versus the exact operational acceptance tasks remaining before full live production sign-off.

Why It Exists

Prevents premature declaration of production readiness and maintains strict change-control transparency.

When to Use:Consult for project sign-off, stakeholder reporting, and release governance.
How to Use:Review the 'Completed in Deployment' checklist alongside the 'Remaining Operational Acceptance' roadmap.
Expected Outcome:Clear, factual understanding of platform maturity without speculative claims.
Important Operational Invariants & Rules:
  • COMPLETED IN THIS RELEASE: Application Code Complete, Website UX Complete, Scanner Accelerated, Discover First-Class Intraday, Insights Calibration Truth, Control Plane Active, Legacy Scheduler Masked/Off, READ_ONLY Active, Broker Disabled, Telegram Authoritative Dispatch LIMITED_LIVE (Legacy CP Adapter DRY_RUN).
  • REMAINING OPERATIONAL ACCEPTANCE (17 Steps): Whole-universe product audit, surface audit, Telegram test recipient commissioning, Canary message test, delivery audit inspection, ordinary recipient verification, admin mirror proof, system routing isolation, LIMITED_LIVE promotion, LIVE promotion, critical backup, restore verification, off-host DR copy, next real trading session acceptance (5m, 15m, 1h, 4h), autonomous post-close EOD acceptance, publication rollover, final freeze.
View-Only Scope:Status ledger is read-only documentation.
Safe to Change:Track progress through the remaining operational roadmap.
Never Touch:Do NOT mark operational steps complete before they are actually executed and proven in live market sessions!
βœ“ Healthy Normal State:All software and UX gates complete; operational steps clearly tracked and scheduled.
⚠️ When Something Is Wrong:Prematurely declaring project complete before live session verification.
Example: Application UX is 100% complete; operational acceptance remains pending next live session at 09:15 IST.
Next: Proceed with scheduled operational acceptance tasks during the next live market session.Open System Cockpit
#what is still pending#production readiness#closeout#remaining acceptance#roadmap#operational sign off#checklist