Tennr · Implementation Engineer · Take-Home
Extract Year, Winner, Score, Runners-up from the first 10 rows of Wikipedia's List of FIFA World Cup finals and append each to Google Sheets via the Sheets API — designed using only the 9 steps in the provided IE Step Library.

▶ View the flowchart in your browser  ·  PNG  ·  PDF

A single-page flowchart that takes an RPA bot from opening the Wikipedia page to one POST per row into Google Sheets. It uses six of the nine library steps — Open page → Loop ×10 { ExtractHTML ×4 · Parse number · Detour · Call API } — and nothing else. Every DOM path and every data value on the chart was verified in a live browser (40/40 field checks), and the three ways Wikipedia's markup has silently drifted since the assignment's resource video was recorded are caught and corrected.

Does this need to be deployed anywhere? No. The deliverable is a static document (a flowchart), not a running service — the assignment explicitly asks for "the flowchart only, no code." The GitHub Pages link above is a convenience so a reviewer can read it in a browser without downloading; it requires zero maintenance. To submit: reply to the assignment email with exports/flowchart.png and exports/flowchart.pdf attached (and, optionally, the live link).

START
 │
 └─ 1  Open page ·············· https://en.wikipedia.org/wiki/List_of_FIFA_World_Cup_finals
 │
 └─ 2  Loop ×10  (index i = 1…10  ↔  table row tbody/tr[i+1]  ↔  sheet row i+1)
 │     │
 │     ├─ 2.1  ExtractHTML   Year        …/tr[{{index}}+1]/th/a   → "1930"
 │     ├─ 2.2  Parse number  "1930" → 1930   (row-validity gate; skips a bad row)
 │     ├─ 2.3  ExtractHTML   Winner      …/tr[{{index}}+1]/td[1]  → "Uruguay"
 │     ├─ 2.4  ExtractHTML   Score       …/tr[{{index}}+1]/td[2]  → "4–2"
 │     ├─ 2.5  ExtractHTML   Runners-up  …/tr[{{index}}+1]/td[3]  → "Argentina"
 │     ├─ 2.6  Detour        all three text fields non-empty?
 │     └─ 2.7  Call API      POST …/values/Sheet1!A:D:append   (one row per iteration)
 │
 └─ 3  Call API  (optional QA) ·· GET the values back, expect 11 rows (header + 10)
 │
END  →  Sheet1!A2:D11 holds the 1930–1974 finals

1 · Verified in a live browser, not assumed. Both selector forms — XPath and CSS — were evaluated against every one of the 10 target rows in a live Chrome session: 80/80 checks returned the expected text. The chart's expected-result table and per-iteration execution trace show real measured values, not placeholders.

2 · Three silent breakages in the resource-video recipe, caught. Since the assignment's DOM-path video was recorded:

• the table prefix drifted from //*[@id="mw-content-text"]/div[1]/table[4] to …/div[2]/section[2]/table[3] — the old path now resolves to nothing;
• the sortable header no longer moves into <thead>, so data row i is tr[i+1], not tr[i];
• the country link's nesting varies row to row (td[1]/a on 1930 but td[1]/span/a on 1954), so the workflow extracts whole cells — the only paths uniform across all 10 rows.

3 · Provable library conformance. A conformance table in the chart quotes each step's Required Input / Output verbatim from the Step Guide next to how the workflow uses it — so "only uses the 9 steps" is verifiable at a glance, not a claim to trust.

4 · Honest about the Guide's own ambiguities. The Guide's Parse number row contradicts itself (Output column says String, description says Int); the chart flags it and makes the step's real job the validity gate, so the design holds under either reading.

POST https://sheets.googleapis.com/v4/spreadsheets/{{SPREADSHEET_ID}}/values/Sheet1!A:D:append
     ?valueInputOption=RAW&insertDataOption=INSERT_ROWS
Authorization: Bearer {{ACCESS_TOKEN}}     # Postman OAuth 2.0, scope .../auth/spreadsheets
Content-Type:  application/json

{ "range": "Sheet1!A:D", "majorDimension": "ROWS",
  "values": [[ 1930, "Uruguay", "4–2", "Argentina" ]] }        # → 200, updatedRange "Sheet1!A2:D2"

RAW stores values verbatim (so an en-dash score is never coerced to a date); INSERT_ROWS only ever adds rows. One POST per iteration pairs with the Loop's skip-on-error rule — a single failed row can't block the other nine. Full path variables, query params, headers, auth config, and failure modes are in panel D of the chart.

(The first 10 rows span 1930–1974 — Wikipedia lists no rows for 1942/1946, cancelled for WWII, so a fixed loop bound of 10 is exact.)

The chart is a single self-contained HTML file; the PNG and PDF are rendered from it with headless Chrome.

chrome --headless --window-size=1160,8400 --force-device-scale-factor=2 \
       --screenshot=exports/flowchart.png  flowchart.html
chrome --headless --no-pdf-header-footer --print-to-pdf=exports/flowchart.pdf  flowchart.html

ten/
├─ flowchart.html          ← source (self-contained HTML)
├─ index.html              ← redirect for GitHub Pages
├─ README.md
└─ exports/
   ├─ flowchart.png        ← primary deliverable
   ├─ flowchart.pdf        ← A3 PDF
   └─ preview-hero.png     ← README banner

_{Bhunesh Bansal · Implementation Engineer take-home · Tennr}