Skip to content

Correction of analyst target data gauge and addition of the 52-Week Range gauge

Ticket #338: Correction of analyst target data gauge and addition of the 52-Week Range gauge
Type: Fix / Business Logic / Test Lockdown
Affected Component: code_source_simule/flask_app.py, templates/titre_detail.html, tests/test_titre_detail.py, docs/tests/stock_detail.md, docs/llm-wiki/01_architecture/key-decisions.md

1. Context and Symptoms

Some data on the stock detail page was problematic: the analyst target gauge reused the 52-week low/high values when the active CSV row did not contain any analyst target data. In plain terms, the screen replaced a missing value with another set of data. The result was misleading for users.

This problem appeared for tickers such as AFLYY and UA, where the CSV contains blank or N/A target values. Instead of an explicit unavailable state, the interface displayed 52-week values from the wrong section.

2. Objective

Keep these two data families separate and clear:

  • analyst targets must read only the target columns from the active CSV;
  • the 52-Week Range must use its own low/high thresholds over 52 weeks;
  • a missing analyst target must remain unavailable and never be replaced by 52-week data;
  • the current price remains shared between the two gauges to keep the reading simple and consistent.

3. Implemented Solution

3.1 — Source-driven analyst lookup

The detail route was updated to resolve analyst targets directly from the active CSV path and to ignore an_haut / an_bas for this gauge. The helper returns None if the CSV row is empty or N/A, which respects the product rule for missing data.

3.2 — Dedicated 52-week range gauge

The 52-Week Range block now has its own gauge, in the same visual style as the analyst gauge but with the labels:

  • 52-Week Low
  • 52-Week High

The comparison continues to use the same current price as the analyst module so the user keeps one consistent reference point.

3.3 — Regression lock with tests

A dedicated test suite was added to lock this business contract:

  • ready-state analyst data from the CSV;
  • unavailable state when CSV values are blank or N/A;
  • clear message when the range is invalid (low > high);
  • regression preventing 52-week thresholds from replacing a missing analyst target;
  • 52-week gauge using the current price and its own low/high thresholds.

4. Validation

The full project suite was executed after the fix:

  • full suite: 268 tests passed;
  • warnings: 3 (existing pandas/pyarrow deprecation plus parser warnings on a known corrupted CSV fixture);
  • application code coverage (code_source_simule/*): 79%.

This validation confirms that the business fix is coherent and that the regression suite remains green.

5. Outcome

The user experience is now aligned with the data contract:

  • missing analyst targets remain missing;
  • 52-week thresholds remain in their own module;
  • both gauges remain readable and comparable with the same current price.

6. Token Consumption Summary

During this session of correction and improvement of the analysis gauges, which lasted more than 2 hours, 175.27 Copilot AI credits were consumed, for a gross cost of $1.75; which represents only 0.7% of the total cost of a consultant on such a task.