Rework Dockerfile.scratch (#2525)

### What problem does this PR solve?

Rework Dockerfile.scratch
- Multiple stage Dockerfile
- Removed conda
- Replaced pip with poetry
- Added missing dependencies and fixed package version conflicts
- Added deepdoc models

### Type of change

- [x] Refactoring
- [ ] Performance Improvement
- [ ] Other (please describe):

requirements.txt

@@ -4,7 +4,7 @@ azure-identity==1.17.1
 azure-storage-file-datalake==12.16.0
 anthropic===0.34.1
 arxiv==2.1.3
-Aspose.Slides==24.2.0
+Aspose.Slides==24.8.0
 BCEmbedding==0.1.3
 Bio==1.7.1
 boto3==1.34.140
@@ -43,7 +43,7 @@ Markdown==3.6
 markdown_to_json==2.1.1
 minio==7.2.4
 mistralai==0.4.2
-nltk==3.9
+nltk==3.9.1
 numpy==1.26.4
 ollama==0.2.1
 onnxruntime==1.17.3
@@ -57,7 +57,6 @@ pandas==2.2.2
 pdfplumber==0.10.4
 peewee==3.17.1
 Pillow==10.3.0
-pipreqs==0.5.0
 protobuf==5.27.2
 psycopg2-binary==2.9.9
 pyclipper==1.3.0.post5