docs: Add high-level Python API documentation [AI-assisted] (#215)

Document the two main public API functions:
- assess_journal() for single journal assessment
- assess_bibtex_file() for batch BibTeX processing

Intentionally minimal to avoid locking in internal APIs.

Fixes #206

Co-authored-by: florath-ai-assistant[bot] <Andreas.Florath@telekom.de>

Author: coding-ai-assistant[bot]

docs/api-reference/python-api.md

@@ -0,0 +1,86 @@
+# Python API Reference
+
+This document describes the high-level public API for programmatic journal assessment.
+
+## Single Journal Assessment
+
+**Function**: `query_dispatcher.assess_journal(query_input)`
+
+Assesses a single journal or conference by querying multiple data sources and returning a consolidated classification.
+
+**Usage**:
+```python
+import asyncio
+from aletheia_probe import query_dispatcher
+from aletheia_probe.normalizer import input_normalizer
+
+async def assess():
+    # Normalize the journal name
+    query = input_normalizer.normalize("Nature Communications")
+
+    # Get assessment
+    result = await query_dispatcher.assess_journal(query)
+
+    # Use results
+    print(f"Classification: {result.assessment}")
+    print(f"Confidence: {result.confidence:.0%}")
+
+    return result
+
+asyncio.run(assess())
+```
+
+**Result object** contains:
+- `assessment`: Classification (predatory, legitimate, suspicious, unknown)
+- `confidence`: Confidence score (0.0 to 1.0)
+- `reasoning`: List of explanation strings
+- `backend_results`: Details from each data source queried
+
+---
+
+## BibTeX File Assessment
+
+**Function**: `BibtexBatchAssessor.assess_bibtex_file(file_path, verbose=False)`
+
+Processes an entire BibTeX file and assesses all journals and conferences referenced.
+
+**Usage**:
+```python
+import asyncio
+from pathlib import Path
+from aletheia_probe import BibtexBatchAssessor
+
+async def assess_bibliography():
+    result = await BibtexBatchAssessor.assess_bibtex_file(
+        Path("references.bib"),
+        verbose=True
+    )
+
+    # Get statistics
+    print(f"Total venues: {result.entries_with_journals}")
+    print(f"Predatory: {result.predatory_count}")
+    print(f"Retracted articles: {result.retracted_articles_count}")
+
+    # Format summary
+    summary = BibtexBatchAssessor.format_summary(result)
+    print(summary)
+
+    return result
+
+asyncio.run(assess_bibliography())
+```
+
+**Result object** contains:
+- `entries_with_journals`: Number of venues assessed
+- `predatory_count`, `legitimate_count`, `suspicious_count`: Classification counts
+- `retracted_articles_count`: Articles found to be retracted
+- `assessment_results`: List of (entry, assessment) tuples for detailed processing
+
+---
+
+## Notes
+
+- Both functions are asynchronous and require `await` or `asyncio.run()`
+- Input normalization handles abbreviations, special characters, and acronyms
+- Results include detailed reasoning for transparency
+- BibTeX assessor automatically detects and skips preprints (arXiv, bioRxiv, etc.)