email-origin-chain 1.0.0

package/docs/walkthrough_address_fix.md

@@ -0,0 +1,58 @@
+# Address Normalization Fix
+
+## Summary
+Fixed critical issue where email addresses from MIME headers were not being properly cleaned, causing test failures. The `normalizeFrom` function now correctly handles complex Outlook address formats with nested `<mailto:...>` patterns.
+
+## Changes Made
+
+### [`src/utils.ts`](file:///c:/Users/Flo/.gemini/antigravity/ProjetsPerso/email-deepest-forward/src/utils.ts)
+- **Rewrote `normalizeFrom` function** with preprocessing step
+  - Strips all `<mailto:...>` patterns using global regex
+  - Counts `<` and `>` brackets to remove excess trailing `>`
+  - Handles complex patterns like: `"Flo M." <florian.mezy@gmail.com<mailto:florian.mezy@gmail.com>>`
+  - Extracts email from `"Name" <email>` format when it appears in address field
+  - Validates extracted emails before returning
+
+### [`src/index.ts`](file:///c:/Users/Flo/.gemini/antigravity/ProjetsPerso/email-deepest-forward/src/index.ts)
+- **Added `normalizeFrom` import** and calls in 3 locations:
+  1. Line 44: Normalize `inlineResult.from` before using it
+  2. Line 53: Normalize MIME metadata fallback `from`
+  3. Line 67: Normalize history root entry `from`
+- **Fixed object spread order**: Destructured `inlineResult` to exclude `from` field, preventing it from overwriting our normalized value
+
+### [`tests/utils.test.ts`](file:///c:/Users/Flo/.gemini/antigravity/ProjetsPerso/email-deepest-forward/tests/utils.test.ts) [NEW]
+- Created comprehensive unit tests for `normalizeFrom`
+- Tests cover:
+  - Standard `Name <email>` format
+  - `email [email]` pattern
+  - `email<mailto:email>` pattern
+  - `email<mailto:email>>` pattern (double bracket)
+  - Email extraction from name field
+  - Edge cases (null, empty, simple addresses)
+
+## Test Results
+
+### ✅ Passing (20/22)
+- All `utils.test.ts` tests pass (8/8)
+- All `crisp-fixtures.test.ts` tests pass
+- All `basic.test.ts`, `comprehensive.test.ts`, `integration.test.ts` pass
+- `empty-body-forward-anonymized.eml` test passes
+
+### ⚠️ Remaining Failures (2/22)
+Both failures are **depth detection issues** (the original problem):
+
+1. **`complex-forward.eml`**
+   - Expected depth: 5
+   - Received depth: 4
+   - Missing 1 forward level detection
+
+2. **`extreme-forward-anonymized.eml`**
+   - Expected depth: 4
+   - Received depth: 3
+   - Missing 1 forward level detection
+
+## Next Steps
+The address normalization is now working correctly. The remaining work is to:
+1. Debug why `complex-forward.eml` only detects 4 levels instead of 5
+2. Debug why `extreme-forward-anonymized.eml` only detects 3 levels instead of 4
+3. This likely involves examining the detector chain and regex patterns in `OutlookEmptyHeaderDetector` and `OutlookReverseFrDetector`

package/docs/walkthrough_deep_forward_fix.md

@@ -0,0 +1,35 @@
+# Deep Forward Parsing Fix - Final Report
+
+## Summary
+Successfully resolved all test failures in `eml-fixture.test.ts`, achieving a **100% pass rate (23/23 tests)**. The library now correctly detects deep nested forwards even in complex mixed scenarios (Outlook, English, French, Empty Headers).
+
+## Key Fixes Implemented
+
+### 1. Fix: Missing Outlook Forward (Empty Header)
+**Problem**: The `OutlookEmptyHeaderDetector` was failing to detect blocks like:
+```text
+________________________________
+De: Florian M.
+```
+**Root Cause**: The regex `(?:_{30,}\s*)?` was "greedy" and consumed the newline after the separator. The subsequent `\r?\nDe` failed because it expected *another* newline which wasn't there.
+**Solution**: Modified `src/detectors/outlook-empty-header-detector.ts` to use `[ \t]*` instead of `\s*` to avoid eating newlines, and improved stability against QP encoding errors.
+
+### 2. Fix: Missing "Intermediate" Levels (Registry Priority)
+**Problem**: In mixed chains, standard detectors (Crisp, priority 0) were detecting *later* forwards first, effectively "skipping" earlier non-standard Outlook forwards.
+**Root Cause**: `DetectorRegistry` returned the *first found* result based on priority, regardless of position in text.
+**Solution**: Updated `src/detectors/registry.ts` to run ALL detectors and return the match that appears **earliest** in the text (smallest index). This ensures we process the conversation chain in correct chronological order.
+
+### 3. Fix: Truncated Body Text ("Florian" issue)
+**Problem**: `OutlookReverseFrDetector` was returning only "Florian" instead of "Genial Yodjii... Florian".
+**Root Cause**: The detector relied on finding a double newline (`\n\n`) to separate headers from body. In some Outlook blocks, the body starts immediately after the `Objet:` line with a single newline. The heuristic skipped the entire message body until it hit a double newline in the signature.
+**Solution**: Updated `src/detectors/outlook-reverse-fr-detector.ts` to assume the body starts immediately after the `Objet` line if no clear separator is found.
+
+## Verified Results
+
+| Fixture | Expected Depth | Received Depth | Status |
+| :--- | :--- | :--- | :--- |
+| `complex-forward.eml` | 5 | 5 | ✅ PASS |
+| `extreme-forward.eml` | 4 | 4 | ✅ PASS |
+| `empty-body...eml` | 2 | 2 | ✅ PASS |
+
+All other tests (`basic`, `utils`, `crisp`) remain passing.

package/package.json

@@ -0,0 +1,48 @@
+{
+  "name": "email-origin-chain",
+  "version": "1.0.0",
+  "description": "Uncover the full audit trail of your email threads. Recursively reconstructs the entire conversation history with instant access to the original sender and true source message.",
+  "main": "dist/index.js",
+  "types": "dist/index.d.ts",
+  "files": [
+    "dist",
+    "README.md",
+    "LICENSE",
+    "docs"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/yodjii/email-origin-chain.git"
+  },
+  "bugs": {
+    "url": "https://github.com/yodjii/email-origin-chain/issues"
+  },
+  "homepage": "https://github.com/yodjii/email-origin-chain#readme",
+  "scripts": {
+    "build": "tsc",
+    "test": "jest",
+    "test:watch": "jest --watch",
+    "prepublishOnly": "npm run build"
+  },
+  "keywords": [
+    "email",
+    "parser",
+    "forward",
+    "mailparser"
+  ],
+  "author": "Flo (yodjii)",
+  "license": "MIT",
+  "devDependencies": {
+    "@types/jest": "^30.0.0",
+    "@types/mailparser": "^3.4.6",
+    "@types/node": "^25.0.10",
+    "jest": "^30.2.0",
+    "ts-jest": "^29.4.6",
+    "typescript": "^5.9.3"
+  },
+  "dependencies": {
+    "any-date-parser": "^2.2.3",
+    "email-forward-parser": "^1.7.2",
+    "mailparser": "^3.9.1"
+  }
+}