docintel-platform 1.2.0__tar.gz → 1.3.0__tar.gz

{docintel_platform-1.2.0/src/docintel_platform.egg-info → docintel_platform-1.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docintel-platform
-Version: 1.2.0
+Version: 1.3.0
 Summary: Enterprise document intelligence API for PDF compliance, multi-format extraction, structuring, and summarization.
 Author: Babandeep Singh
 License-Expression: MIT
@@ -40,9 +40,11 @@ Requires-Dist: fakeredis>=2.26.2; extra == "dev"
 Requires-Dist: prometheus-client>=0.21.0; extra == "dev"
 Requires-Dist: python-docx>=1.1.2; extra == "dev"
 Requires-Dist: openpyxl>=3.1.5; extra == "dev"
+Requires-Dist: python-pptx>=1.0.2; extra == "dev"
 Provides-Extra: documents
 Requires-Dist: python-docx>=1.1.2; extra == "documents"
 Requires-Dist: openpyxl>=3.1.5; extra == "documents"
+Requires-Dist: python-pptx>=1.0.2; extra == "documents"
 Provides-Extra: ocr
 Requires-Dist: easyocr>=1.7.2; extra == "ocr"
 Requires-Dist: presidio-analyzer>=2.2.354; extra == "ocr"
@@ -80,6 +82,7 @@ Requires-Dist: cryptography>=43.0.0; extra == "all"
 Requires-Dist: gradio>=4.44.0; extra == "all"
 Requires-Dist: python-docx>=1.1.2; extra == "all"
 Requires-Dist: openpyxl>=3.1.5; extra == "all"
+Requires-Dist: python-pptx>=1.0.2; extra == "all"
 
 # Document Intelligence Platform
 
@@ -90,7 +93,7 @@ Requires-Dist: openpyxl>=3.1.5; extra == "all"
 
 Enterprise document intelligence API: PDF compliance (OCR, PII, redaction), LLM structuring, and multi-format text workflows (Word, Excel, CSV, plain text).
 
-**Version:** 1.2.0 | **PyPI:** [docintel-platform](https://pypi.org/project/docintel-platform/)
+**Version:** 1.3.0 | **PyPI:** [docintel-platform](https://pypi.org/project/docintel-platform/)
 
 ---
 
@@ -119,7 +122,7 @@ Gradio includes a **Document process** tab (unified pipeline). It needs the API
 ```bash
 pip install docintel-platform
 pip install "docintel-platform[all]"        # OCR, LLM, jobs, auth, UI, office formats
-pip install "docintel-platform[documents]"  # Word and Excel only
+pip install "docintel-platform[documents]"  # Word, Excel, and PowerPoint
 ```
 
 **Python client:**
@@ -141,13 +144,13 @@ report = client.process_document("policy.docx", include_pii=True)
 | PDF annotate | `POST /v1/pdf/annotate` | Regex highlight, redact, markup |
 | PDF PII scan | `POST /v1/pdf/detect-sensitive` | Presidio + OCR for scanned PDFs |
 | PDF structure | `POST /v1/pdf/structure` | OCR + LLM curated PDF (needs LLM key) |
-| Documents | `POST /v1/documents/*` | Identify, extract, classify, summarize, PII, compare, **process** |
+| Documents | `POST /v1/documents/*` | Identify, extract, classify, summarize, PII, compare, **process**, **ingest** (S3) |
 | Text | `POST /v1/text/summarize` | TextRank extractive summary |
 | Batch | `POST /v1/batch` | Async summarize, classify, detect_pii, process |
 | Jobs | `GET /v1/jobs/{id}` | Poll async work (`?async=true`; default in Docker when Redis is up) |
 | Ops | `GET /health`, `GET /metrics` | Health and Prometheus-friendly metrics |
 
-**Supported uploads (text workflows):** PDF, DOCX, XLSX, CSV, JSON, TXT, MD.
+**Supported uploads (text workflows):** PDF, DOCX, XLSX, PPTX, CSV, JSON, TXT, MD.
 
 **PDF-only routes** (annotate, sensitive, structure) return HTTP 415 for other types. Use `/v1/documents/extract-text` or `/v1/documents/process` for office files.
 
@@ -199,6 +202,7 @@ Copy `.env.example` to `.env` for `DOCINTEL_LLM_API_KEY`, auth keys, Redis, and
 | [docs/PLATFORM.md](docs/PLATFORM.md) | Jobs, auth, storage, ops layout |
 | [docs/PRODUCTION.md](docs/PRODUCTION.md) | Checklist, latency, failure modes |
 | [docs/ROADMAP.md](docs/ROADMAP.md) | Milestones and history |
+| [docs/WEBHOOKS.md](docs/WEBHOOKS.md) | Async callbacks and S3 ingest |
 | [docs/adr/](docs/adr/) | Architecture decision records |
 
 ---

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/README.md

@@ -7,7 +7,7 @@
 
 Enterprise document intelligence API: PDF compliance (OCR, PII, redaction), LLM structuring, and multi-format text workflows (Word, Excel, CSV, plain text).
 
-**Version:** 1.2.0 | **PyPI:** [docintel-platform](https://pypi.org/project/docintel-platform/)
+**Version:** 1.3.0 | **PyPI:** [docintel-platform](https://pypi.org/project/docintel-platform/)
 
 ---
 
@@ -36,7 +36,7 @@ Gradio includes a **Document process** tab (unified pipeline). It needs the API
 ```bash
 pip install docintel-platform
 pip install "docintel-platform[all]"        # OCR, LLM, jobs, auth, UI, office formats
-pip install "docintel-platform[documents]"  # Word and Excel only
+pip install "docintel-platform[documents]"  # Word, Excel, and PowerPoint
 ```
 
 **Python client:**
@@ -58,13 +58,13 @@ report = client.process_document("policy.docx", include_pii=True)
 | PDF annotate | `POST /v1/pdf/annotate` | Regex highlight, redact, markup |
 | PDF PII scan | `POST /v1/pdf/detect-sensitive` | Presidio + OCR for scanned PDFs |
 | PDF structure | `POST /v1/pdf/structure` | OCR + LLM curated PDF (needs LLM key) |
-| Documents | `POST /v1/documents/*` | Identify, extract, classify, summarize, PII, compare, **process** |
+| Documents | `POST /v1/documents/*` | Identify, extract, classify, summarize, PII, compare, **process**, **ingest** (S3) |
 | Text | `POST /v1/text/summarize` | TextRank extractive summary |
 | Batch | `POST /v1/batch` | Async summarize, classify, detect_pii, process |
 | Jobs | `GET /v1/jobs/{id}` | Poll async work (`?async=true`; default in Docker when Redis is up) |
 | Ops | `GET /health`, `GET /metrics` | Health and Prometheus-friendly metrics |
 
-**Supported uploads (text workflows):** PDF, DOCX, XLSX, CSV, JSON, TXT, MD.
+**Supported uploads (text workflows):** PDF, DOCX, XLSX, PPTX, CSV, JSON, TXT, MD.
 
 **PDF-only routes** (annotate, sensitive, structure) return HTTP 415 for other types. Use `/v1/documents/extract-text` or `/v1/documents/process` for office files.
 
@@ -116,6 +116,7 @@ Copy `.env.example` to `.env` for `DOCINTEL_LLM_API_KEY`, auth keys, Redis, and
 | [docs/PLATFORM.md](docs/PLATFORM.md) | Jobs, auth, storage, ops layout |
 | [docs/PRODUCTION.md](docs/PRODUCTION.md) | Checklist, latency, failure modes |
 | [docs/ROADMAP.md](docs/ROADMAP.md) | Milestones and history |
+| [docs/WEBHOOKS.md](docs/WEBHOOKS.md) | Async callbacks and S3 ingest |
 | [docs/adr/](docs/adr/) | Architecture decision records |
 
 ---

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "docintel-platform"
-version = "1.2.0"
+version = "1.3.0"
 description = "Enterprise document intelligence API for PDF compliance, multi-format extraction, structuring, and summarization."
 readme = "README.md"
 license = "MIT"
@@ -50,10 +50,12 @@ dev = [
     "prometheus-client>=0.21.0",
     "python-docx>=1.1.2",
     "openpyxl>=3.1.5",
+    "python-pptx>=1.0.2",
 ]
 documents = [
     "python-docx>=1.1.2",
     "openpyxl>=3.1.5",
+    "python-pptx>=1.0.2",
 ]
 ocr = [
     "easyocr>=1.7.2",
@@ -98,6 +100,7 @@ all = [
     "gradio>=4.44.0",
     "python-docx>=1.1.2",
     "openpyxl>=3.1.5",
+    "python-pptx>=1.0.2",
 ]
 
 [project.scripts]

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/__init__.py

@@ -2,5 +2,5 @@
 
 from docintel.client import DocintelClient, DocintelError
 
-__version__ = "1.2.0"
+__version__ = "1.3.0"
 __all__ = ["DocintelClient", "DocintelError", "__version__"]

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/capabilities/extraction/formats/extract.py

@@ -39,6 +39,8 @@ def extract_document_text(
         return _extract_docx(file_path, resolved)
     if resolved.kind is DocumentKind.XLSX:
         return _extract_xlsx(file_path, resolved)
+    if resolved.kind is DocumentKind.PPTX:
+        return _extract_pptx(file_path, resolved)
     if resolved.kind is DocumentKind.CSV:
         return _extract_csv(file_path, resolved)
     if resolved.kind is DocumentKind.JSON:
@@ -129,6 +131,37 @@ def _extract_xlsx(path: Path, identification: IdentificationResult) -> Extractio
     )
 
 
+def _extract_pptx(path: Path, identification: IdentificationResult) -> ExtractionResult:
+    try:
+        from pptx import Presentation
+    except ImportError as exc:
+        raise RuntimeError(
+            "PowerPoint support requires optional dependencies. Install: pip install -e '.[documents]'"
+        ) from exc
+
+    presentation = Presentation(path)
+    segments: list[dict] = []
+    parts: list[str] = []
+    for slide_index, slide in enumerate(presentation.slides, start=1):
+        slide_parts: list[str] = []
+        for shape in slide.shapes:
+            text = getattr(shape, "text", "").strip()
+            if text:
+                slide_parts.append(text)
+        slide_text = "\n".join(slide_parts)
+        segments.append({"slide": slide_index, "text": slide_text})
+        if slide_text:
+            parts.append(f"# Slide {slide_index}\n{slide_text}")
+
+    return ExtractionResult(
+        kind=identification.kind,
+        mime_type=identification.mime_type,
+        text="\n\n".join(parts),
+        segments=segments,
+        metadata={"slide_count": len(presentation.slides)},
+    )
+
+
 def _extract_csv(path: Path, identification: IdentificationResult) -> ExtractionResult:
     raw = path.read_text(encoding="utf-8", errors="replace")
     sample = raw[:2048]

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/capabilities/extraction/formats/models.py

@@ -11,6 +11,7 @@ class DocumentKind(str, Enum):
     PDF = "pdf"
     DOCX = "docx"
     XLSX = "xlsx"
+    PPTX = "pptx"
     CSV = "csv"
     PLAIN_TEXT = "plain_text"
     JSON = "json"

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/capabilities/extraction/formats/registry.py

@@ -31,6 +31,14 @@ _PROFILES: tuple[DocumentProfile, ...] = (
         supports_pdf_pipeline=False,
         supports_text_extraction=True,
     ),
+    DocumentProfile(
+        kind=DocumentKind.PPTX,
+        mime_type="application/vnd.openxmlformats-officedocument.presentationml.presentation",
+        extensions=(".pptx",),
+        label="PowerPoint presentation",
+        supports_pdf_pipeline=False,
+        supports_text_extraction=True,
+    ),
     DocumentProfile(
         kind=DocumentKind.CSV,
         mime_type="text/csv",

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/capabilities/extraction/formats/sniff.py

@@ -31,6 +31,8 @@ def _sniff_zip_kind(path: Path) -> DocumentKind | None:
 
     if any(name.startswith("word/") for name in names):
         return DocumentKind.DOCX
+    if any(name.startswith("ppt/") for name in names):
+        return DocumentKind.PPTX
     if any(name.startswith("xl/") for name in names):
         return DocumentKind.XLSX
     return None
@@ -80,7 +82,7 @@ def _looks_like_csv(sample: str) -> bool:
 
 
 def _requires_content_confirmation(kind: DocumentKind) -> bool:
-    return kind in {DocumentKind.PDF, DocumentKind.DOCX, DocumentKind.XLSX}
+    return kind in {DocumentKind.PDF, DocumentKind.DOCX, DocumentKind.XLSX, DocumentKind.PPTX}
 
 
 def _build_result(

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/client.py

@@ -416,3 +416,45 @@ class DocintelClient:
                 data=data,
                 poll=poll,
             )
+
+    def ingest_document_from_s3(
+        self,
+        *,
+        s3_uri: str | None = None,
+        bucket: str | None = None,
+        key: str | None = None,
+        sentences: int = 3,
+        include_summarize: bool = True,
+        include_pii: bool = True,
+        include_text: bool = False,
+        entities: str | None = None,
+        vertical: str | None = None,
+        min_score: float = 0.35,
+        callback_url: str | None = None,
+        poll: bool = True,
+    ) -> dict[str, Any]:
+        body: dict[str, Any] = {
+            "operation": "process",
+            "sentences": sentences,
+            "include_summarize": include_summarize,
+            "include_pii": include_pii,
+            "include_text": include_text,
+            "min_score": min_score,
+        }
+        if s3_uri:
+            body["s3_uri"] = s3_uri
+        if bucket:
+            body["bucket"] = bucket
+        if key:
+            body["key"] = key
+        if entities:
+            body["entities"] = entities
+        if vertical:
+            body["vertical"] = vertical
+        if callback_url:
+            body["callback_url"] = callback_url
+        return self._post_async_json(
+            "/v1/documents/ingest",
+            json_body=body,
+            poll=poll,
+        )

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/config.py

@@ -11,6 +11,7 @@ class Config:
     LOG_LEVEL = os.getenv("DOCINTEL_LOG_LEVEL", "INFO")
     REDIS_URL = os.getenv("DOCINTEL_REDIS_URL", "redis://localhost:6379/0")
     JOBS_ENABLED = os.getenv("DOCINTEL_JOBS_ENABLED", "true").lower() == "true"
+    JOB_TTL_SECONDS = int(os.getenv("DOCINTEL_JOB_TTL_SECONDS", str(60 * 60 * 24 * 7)))
     QUEUE_NAME = os.getenv("DOCINTEL_QUEUE_NAME", "docintel")
     API_KEYS = os.getenv("DOCINTEL_API_KEYS", "")
     AUTH_REQUIRED = os.getenv("DOCINTEL_AUTH_REQUIRED", "false").lower() == "true"

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/jobs/models.py

@@ -35,6 +35,7 @@ class JobType(str, Enum):
     DOCUMENT_DETECT_PII = "document_detect_pii"
     DOCUMENT_EXTRACT_TEXT = "document_extract_text"
     DOCUMENT_COMPARE = "document_compare"
+    DOCUMENT_S3_PROCESS = "document_s3_process"
     BATCH = "batch"
 
 

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/jobs/queue.py

@@ -270,6 +270,25 @@ def enqueue_extract_text_job(
     )
 
 
+def enqueue_s3_document_process_job(
+    job_id: str,
+    bucket: str,
+    key: str,
+    options: dict,
+) -> None:
+    queue = get_queue()
+    queue.enqueue(
+        "docintel.jobs.tasks.run_s3_document_process_job",
+        job_id=job_id,
+        bucket=bucket,
+        key=key,
+        options=options,
+        job_timeout=900,
+        result_ttl=DEFAULT_RESULT_TTL,
+        failure_ttl=DEFAULT_FAILURE_TTL,
+    )
+
+
 def enqueue_compare_job(
     job_id: str,
     *,

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/jobs/store.py

@@ -12,6 +12,15 @@ JOB_KEY_PREFIX = "docintel:job:"
 DEFAULT_JOB_TTL_SECONDS = 60 * 60 * 24 * 7
 
 
+def job_ttl_seconds() -> int:
+    raw = os.getenv("DOCINTEL_JOB_TTL_SECONDS", str(DEFAULT_JOB_TTL_SECONDS)).strip()
+    try:
+        ttl = int(raw)
+    except ValueError:
+        return DEFAULT_JOB_TTL_SECONDS
+    return max(ttl, 60)
+
+
 def redis_url() -> str:
     return os.getenv("DOCINTEL_REDIS_URL", "redis://localhost:6379/0").strip()
 
@@ -37,9 +46,10 @@ def _job_key(job_id: str) -> str:
     return f"{JOB_KEY_PREFIX}{job_id}"
 
 
-def save_job(record: JobRecord, ttl_seconds: int = DEFAULT_JOB_TTL_SECONDS) -> None:
+def save_job(record: JobRecord, ttl_seconds: int | None = None) -> None:
     client = _redis_client()
-    client.set(_job_key(record.job_id), json.dumps(record.to_dict()), ex=ttl_seconds)
+    resolved_ttl = ttl_seconds if ttl_seconds is not None else job_ttl_seconds()
+    client.set(_job_key(record.job_id), json.dumps(record.to_dict()), ex=resolved_ttl)
 
 
 def get_job(job_id: str) -> JobRecord | None:

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/jobs/tasks.py

@@ -645,6 +645,45 @@ def run_compare_job(
     )
 
 
+def run_s3_document_process_job(
+    *,
+    job_id: str,
+    bucket: str,
+    key: str,
+    options: dict,
+) -> dict:
+    from docintel.storage.s3_ingest import download_s3_object_to_job_dir
+
+    record = get_job(job_id)
+    callback_url = record.callback_url if record else None
+    update_job(
+        job_id,
+        job_status=JobStatus.RUNNING.value,
+        progress=5,
+        progress_message="Downloading from S3",
+    )
+    try:
+        input_path, filename = download_s3_object_to_job_dir(job_id, bucket, key)
+    except Exception as exc:
+        failed = update_job(
+            job_id,
+            job_status=JobStatus.FAILED.value,
+            progress=100,
+            progress_message="Job failed",
+            error=str(exc),
+        )
+        _notify_webhook(callback_url, failed)
+        raise
+
+    return run_document_process_job(
+        job_id=job_id,
+        input_path=str(input_path),
+        filename=filename,
+        content_type=None,
+        options=options,
+    )
+
+
 def create_queued_job(
     job_id: str,
     *,

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/openapi/openapi.yaml

@@ -541,6 +541,66 @@ paths:
         "503":
           description: Presidio stack unavailable
 
+  /v1/documents/ingest:
+    post:
+      tags: [documents]
+      summary: Queue process pipeline for an S3 object
+      description: |
+        Downloads the object in the worker, then runs the same pipeline as
+        POST /v1/documents/process. Always returns 202 when Redis is available.
+      requestBody:
+        required: true
+        content:
+          application/json:
+            schema:
+              type: object
+              required: [operation]
+              properties:
+                s3_uri:
+                  type: string
+                  example: s3://my-bucket/inbox/policy.docx
+                bucket:
+                  type: string
+                key:
+                  type: string
+                operation:
+                  type: string
+                  enum: [process]
+                  default: process
+                sentences:
+                  type: integer
+                  minimum: 1
+                  maximum: 20
+                include_summarize:
+                  type: boolean
+                  default: true
+                include_pii:
+                  type: boolean
+                  default: true
+                include_text:
+                  type: boolean
+                  default: false
+                vertical:
+                  type: string
+                entities:
+                  type: string
+                min_score:
+                  type: number
+                callback_url:
+                  type: string
+                  format: uri
+      responses:
+        "202":
+          description: S3 ingest job queued
+          content:
+            application/json:
+              schema:
+                $ref: "#/components/schemas/AsyncAccepted"
+        "400":
+          description: Invalid S3 location or options
+        "503":
+          description: Async jobs unavailable
+
   /v1/documents/process:
     post:
       tags: [documents]

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/routes/documents.py

@@ -182,6 +182,60 @@ def _parse_process_options() -> tuple[ProcessOptions | None, dict | None, int |
     )
 
 
+def _parse_process_options_from_dict(
+    payload: dict,
+) -> tuple[ProcessOptions | None, dict | None, int | None]:
+    raw_sentences = payload.get("sentences", DEFAULT_SENTENCE_COUNT)
+    try:
+        sentences = int(raw_sentences)
+    except (TypeError, ValueError):
+        return None, {"error": "Field 'sentences' must be an integer."}, 400
+    if sentences < 1 or sentences > MAX_SENTENCE_COUNT:
+        return None, {
+            "error": f"Field 'sentences' must be between 1 and {MAX_SENTENCE_COUNT}."
+        }, 400
+
+    vertical = payload.get("vertical", "")
+    vertical = vertical.strip() if isinstance(vertical, str) else ""
+    entities_raw = payload.get("entities")
+    try:
+        entities = _resolve_entities(
+            entities_raw if isinstance(entities_raw, str) else None,
+            vertical or None,
+        )
+    except ValueError as exc:
+        return None, {"error": str(exc)}, 400
+
+    raw_min_score = payload.get("min_score", 0.35)
+    try:
+        min_score = float(raw_min_score)
+    except (TypeError, ValueError):
+        return None, {"error": "Field 'min_score' must be a number."}, 400
+
+    def _bool_value(name: str, default: bool) -> bool:
+        if name not in payload:
+            return default
+        value = payload[name]
+        if isinstance(value, bool):
+            return value
+        if isinstance(value, str):
+            return value.strip().lower() in {"1", "true", "yes", "on"}
+        return default
+
+    return (
+        ProcessOptions(
+            sentences=sentences,
+            include_summarize=_bool_value("include_summarize", True),
+            include_pii=_bool_value("include_pii", True),
+            include_text=_bool_value("include_text", False),
+            entities=entities,
+            min_score=min_score,
+        ),
+        None,
+        None,
+    )
+
+
 @documents_bp.get("/types")
 @limiter.limit("120 per hour")
 def supported_document_types():
@@ -189,6 +243,47 @@ def supported_document_types():
     return jsonify({"status": "ok", "types": list_supported_types()})
 
 
+@documents_bp.post("/ingest")
+@limiter.limit("20 per hour")
+def ingest_document():
+    """Queue unified document processing for an object already stored in S3."""
+    payload = request.get_json(silent=True)
+    if not isinstance(payload, dict):
+        return jsonify({"error": "Request body must be JSON."}), 400
+
+    operation = str(payload.get("operation", "process")).strip().lower()
+    if operation != "process":
+        return jsonify({"error": "Only operation 'process' is supported."}), 400
+
+    from docintel.storage.s3_ingest import resolve_s3_location
+
+    try:
+        bucket, key = resolve_s3_location(payload)
+    except ValueError as exc:
+        return jsonify({"error": str(exc)}), 400
+
+    options, option_error, option_status = _parse_process_options_from_dict(payload)
+    if option_error is not None:
+        return jsonify(option_error), option_status
+
+    callback_raw = payload.get("callback_url", "")
+    callback_url = callback_raw.strip() if isinstance(callback_raw, str) and callback_raw.strip() else None
+
+    from docintel.jobs.models import JobType
+    from docintel.jobs.queue import enqueue_s3_document_process_job
+
+    job_id = uuid.uuid4().hex[:12]
+    return enqueue_background_job(
+        job_type=JobType.DOCUMENT_S3_PROCESS,
+        callback_url=callback_url,
+        enqueue_fn=enqueue_s3_document_process_job,
+        job_id=job_id,
+        bucket=bucket,
+        key=key,
+        options=options.to_dict() if options else {},
+    )
+
+
 @documents_bp.post("/identify")
 @limiter.limit("120 per hour")
 def identify_upload():

docintel_platform-1.3.0/src/docintel/storage/s3_ingest.py

@@ -0,0 +1,61 @@
+"""Download objects from S3 for async document ingest."""
+
+from __future__ import annotations
+
+import os
+import re
+from pathlib import Path
+from urllib.parse import unquote
+
+from werkzeug.utils import secure_filename
+
+_S3_URI_PATTERN = re.compile(r"^s3://([^/]+)/(.+)$")
+
+
+def parse_s3_uri(uri: str) -> tuple[str, str]:
+    """Parse s3://bucket/key into bucket and key."""
+    normalized = uri.strip()
+    match = _S3_URI_PATTERN.match(normalized)
+    if not match:
+        raise ValueError("s3_uri must look like s3://bucket/path/to/object")
+    bucket = match.group(1).strip()
+    key = unquote(match.group(2).strip())
+    if not bucket or not key:
+        raise ValueError("s3_uri must include a bucket name and object key")
+    return bucket, key
+
+
+def resolve_s3_location(payload: dict) -> tuple[str, str]:
+    """Resolve bucket and key from JSON body fields."""
+    s3_uri = payload.get("s3_uri")
+    if isinstance(s3_uri, str) and s3_uri.strip():
+        return parse_s3_uri(s3_uri)
+
+    bucket = payload.get("bucket")
+    key = payload.get("key")
+    if isinstance(bucket, str) and bucket.strip() and isinstance(key, str) and key.strip():
+        return bucket.strip(), key.strip()
+
+    raise ValueError("Provide s3_uri or both bucket and key.")
+
+
+def s3_client():
+    import boto3
+
+    return boto3.client(
+        "s3",
+        region_name=os.getenv("DOCINTEL_S3_REGION", "us-east-1"),
+        endpoint_url=os.getenv("DOCINTEL_S3_ENDPOINT_URL", "") or None,
+    )
+
+
+def download_s3_object_to_job_dir(job_id: str, bucket: str, key: str) -> tuple[Path, str]:
+    """Download an S3 object into the job work directory."""
+    from docintel.storage import get_storage
+
+    filename = secure_filename(Path(key).name) or "document.bin"
+    work_dir = get_storage().job_dir(job_id)
+    work_dir.mkdir(parents=True, exist_ok=True)
+    destination = work_dir / filename
+    s3_client().download_file(bucket, key, str(destination))
+    return destination, filename

{docintel_platform-1.2.0 → docintel_platform-1.3.0}/src/docintel/ui.py

@@ -516,12 +516,12 @@ def build_ui():
                 outputs=summary_output,
             )
 
-        office_types = [".pdf", ".docx", ".xlsx", ".csv", ".txt", ".md", ".json"]
+        office_types = [".pdf", ".docx", ".xlsx", ".pptx", ".csv", ".txt", ".md", ".json"]
         with gr.Tab("Document process"):
             gr.Markdown(
                 "Run extract, classify, summarize, and PII detection in one async job. "
                 "Requires Redis and a worker (`make run-worker` or docker-compose worker). "
-                "Word and Excel need `pip install -e '.[documents]'` on the API server."
+                "Office formats need `pip install -e '.[documents]'` on the API server (Word, Excel, PowerPoint)."
             )
             from docintel.capabilities.compliance.presets import list_vertical_presets
 
@@ -563,7 +563,7 @@ def build_ui():
         with gr.Tab("Document tools"):
             gr.Markdown(
                 "Identify, extract, classify, summarize, scan for PII, and compare office documents. "
-                "Requires `pip install -e '.[documents]'` for Word and Excel."
+                "Requires `pip install -e '.[documents]'` for Word, Excel, and PowerPoint."
             )
             with gr.Row():
                 doc_file = gr.File(label="Document upload", file_types=office_types)