sf-translation

You are helping the user generate Salesforce STF translation files. The Salesforce object name is: $ARGUMENTS

If no object name was provided in $ARGUMENTS, immediately ask the user: "Which Salesforce object do you want to generate translations for? (e.g. Vehicle, Case, Account)"

Set OBJECT_NAME = $ARGUMENTS (trimmed).

---

Step 0: Install Dependencies

Run the following silently:

pip3 install -r ~/.claude/skills/sf-translation/requirements.txt -q 2>&1 | tail -3

If it fails, tell the user to run pip3 install pandas openpyxl manually and continue.

---

Step 1: Select Salesforce Org

Run:

python3 ~/.claude/skills/sf-translation/scripts/org_connect.py

Parse the JSON output and display a numbered list to the user:

Authenticated Salesforce orgs:
  1. alias=myorg  username=user@example.com  status=Connected
  2. alias=devorg username=dev@example.com   status=Connected
  0. Connect a new org

Ask: "Enter the number of the org to use:"

If the user picks 0 (new org), run:

sf org login web

Then re-run org_connect.py and show the updated list. Ask the user to select their newly connected org.

Store the selected org's username as SELECTED_ORG.

---

Step 2: Collect File Paths

Ask the user for the following (ask all at once in a single message):

1. Master Excel Sheet path — the translation reference file (Column C = English label, Column D = Spanish, Column E = Portuguese)
2. Output directory — where to save generated files (default: ~/Desktop/sf-translation-output)
3. (Optional) Existing Spanish STF path — a previously downloaded Bilingual STF for Spanish; keys already in the TRANSLATED section will be skipped. Press Enter to skip.
4. (Optional) Existing Portuguese STF path — a previously downloaded Bilingual STF for Portuguese; keys already in the TRANSLATED section will be skipped. Press Enter to skip.

If the user presses Enter for the output directory, use ~/Desktop/sf-translation-output and create it.

Store as: MASTER_PATH, OUTPUT_DIR, EXISTING_ES (empty string if not provided), EXISTING_PT (empty string if not provided).

---

Step 3: Extract Fields from Salesforce

Run:

python3 ~/.claude/skills/sf-translation/scripts/extract_fields.py \
  --org "SELECTED_ORG" \
  --object "OBJECT_NAME" \
  --output "OUTPUT_DIR"

This will:

• Run sf sobject describe against the selected org
• Extract all custom fields and their picklist values
• Generate OUTPUT_DIR/OBJECT_NAME_intermediate.xlsx with two tabs:
  • Custom_Fields: Field Label, Field API Name
  • Picklist_Values: Field API Name, Picklist Value, Picklist Label

Report the printed summary to the user (field count, picklist value count).

If the command fails (non-zero exit), show the error output and stop. Common issues:

• Org not authorized: re-run Step 1
• Object not found: confirm the object API name with the user

---

Step 4: Match Against Master Sheet

Run:

python3 ~/.claude/skills/sf-translation/scripts/compare_master.py \
  --intermediate "OUTPUT_DIR/OBJECT_NAME_intermediate.xlsx" \
  --master "MASTER_PATH" \
  --output "OUTPUT_DIR/OBJECT_NAME_matches.json"

This compares labels in the intermediate Excel against Column C of the Master Sheet and writes a matches JSON file. Report the printed summary (matched count, unmatched count).

---

Step 5: Generate STF Files

Build the command arguments for existing STF files:

• If EXISTING_ES is not empty: add --existing-es "EXISTING_ES"
• If EXISTING_PT is not empty: add --existing-pt "EXISTING_PT"

Run:

python3 ~/.claude/skills/sf-translation/scripts/generate_stf.py \
  --matches "OUTPUT_DIR/OBJECT_NAME_matches.json" \
  --object "OBJECT_NAME" \
  --output "OUTPUT_DIR" \
  [--existing-es "EXISTING_ES"] \
  [--existing-pt "EXISTING_PT"]

This generates:

• OUTPUT_DIR/OBJECT_NAME_es.stf
• OUTPUT_DIR/OBJECT_NAME_pt_BR.stf

Report the printed summary (written count, skipped count per language).

---

Step 6: Generate Miss Report

Run:

python3 ~/.claude/skills/sf-translation/scripts/miss_report.py \
  --matches "OUTPUT_DIR/OBJECT_NAME_matches.json" \
  --object "OBJECT_NAME" \
  --output "OUTPUT_DIR"

This generates OUTPUT_DIR/OBJECT_NAME_miss_report.csv.

---

Final Summary

Tell the user:

Translation files generated for [OBJECT_NAME]:

  OUTPUT_DIR/OBJECT_NAME_intermediate.xlsx  — Field/picklist inventory
  OUTPUT_DIR/OBJECT_NAME_es.stf             — Spanish translations
  OUTPUT_DIR/OBJECT_NAME_pt_BR.stf          — Portuguese (Brazil) translations
  OUTPUT_DIR/OBJECT_NAME_miss_report.csv    — Fields with no translation found

[Print the match/skip/miss counts from each step]