run all dependencies and increase nsjail memory

Author: mbalunovic

Dockerfile

@@ -31,9 +31,10 @@ RUN curl -sSf https://rye.astral.sh/get | RYE_INSTALL_OPTION="--yes" RYE_TOOLCHA
 
 COPY --chown=app requirements.lock pyproject.toml .python-version README.md ./
 
-RUN /bin/bash -c 'source $RYE_HOME/env && UV_NO_CACHE=1 rye sync' && \
+RUN /bin/bash -c 'source $RYE_HOME/env && UV_NO_CACHE=1 rye sync --features=full' && \
     curl https://semgrep.dev/c/r/bash -s -o /tmp/bash && \
     curl https://semgrep.dev/c/r/python.lang.security -s -o /tmp/python.lang.security
+RUN rye run python -m spacy download en_core_web_lg
 
 COPY --chown=app server ./server
 COPY --from=frontend-builder --chown=app /app/dist ./playground/dist

requirements-dev.lock

@@ -3,7 +3,7 @@
 #
 # last locked with the following flags:
 #   pre: false
-#   features: []
+#   features: ["full"]
 #   all-features: false
 #   with-sources: false
 #   generate-hashes: false
@@ -24,6 +24,8 @@ attrs==23.2.0
     # via jsonschema
     # via referencing
     # via semgrep
+blis==0.7.11
+    # via thinc
 boltons==21.0.0
     # via face
     # via glom
@@ -33,6 +35,10 @@ bracex==2.4
 cachetools==5.5.0
     # via asyncache
     # via invariant-server
+catalogue==2.0.10
+    # via spacy
+    # via srsly
+    # via thinc
 certifi==2024.7.4
     # via httpcore
     # via httpx
@@ -46,8 +52,17 @@ click==8.1.7
     # via uvicorn
 click-option-group==0.5.6
     # via semgrep
+cloudpathlib==0.19.0
+    # via weasel
 colorama==0.4.6
     # via semgrep
+confection==0.1.5
+    # via thinc
+    # via weasel
+cymem==2.0.8
+    # via preshed
+    # via spacy
+    # via thinc
 defusedxml==0.7.1
     # via semgrep
 dnspython==2.6.1
@@ -65,6 +80,14 @@ fastapi-cli==0.0.4
     # via fastapi
 fastapi-utils==0.7.0
     # via invariant-server
+filelock==3.15.4
+    # via huggingface-hub
+    # via tldextract
+    # via torch
+    # via transformers
+fsspec==2024.6.1
+    # via huggingface-hub
+    # via torch
 glom==22.1.0
     # via semgrep
 h11==0.14.0
@@ -76,53 +99,96 @@ httptools==0.6.1
     # via uvicorn
 httpx==0.27.0
     # via fastapi
+huggingface-hub==0.24.6
+    # via invariant-server
+    # via tokenizers
+    # via transformers
 idna==3.7
     # via anyio
     # via email-validator
     # via httpx
     # via requests
+    # via tldextract
 iniconfig==2.0.0
     # via pytest
 invariant @ git+https://github.com/invariantlabs-ai/invariant.git@0a5489d3e647858a78f62a8f5abef2b79374ab0b
     # via invariant-server
 jinja2==3.1.4
     # via fastapi
+    # via spacy
+    # via torch
 jsonschema==4.23.0
     # via semgrep
 jsonschema-specifications==2023.12.1
     # via jsonschema
+langcodes==3.4.0
+    # via spacy
+language-data==1.2.0
+    # via langcodes
 lark==1.1.9
     # via invariant
+marisa-trie==1.2.0
+    # via language-data
 markdown-it-py==3.0.0
     # via rich
 markupsafe==2.1.5
     # via jinja2
 mdurl==0.1.2
     # via markdown-it-py
+mpmath==1.3.0
+    # via sympy
+murmurhash==1.0.10
+    # via preshed
+    # via spacy
+    # via thinc
 mypy-extensions==1.0.0
     # via typing-inspect
+networkx==3.3
+    # via torch
+numpy==1.26.4
+    # via blis
+    # via invariant-server
+    # via spacy
+    # via thinc
+    # via transformers
 orjson==3.10.6
     # via fastapi
 packaging==24.1
+    # via huggingface-hub
     # via pytest
     # via semgrep
+    # via spacy
+    # via thinc
+    # via transformers
+    # via weasel
 peewee==3.17.6
     # via semgrep
+phonenumbers==8.13.44
+    # via presidio-analyzer
 pip==24.1.2
     # via invariant
 pluggy==1.5.0
     # via pytest
+preshed==3.0.9
+    # via spacy
+    # via thinc
+presidio-analyzer==2.2.355
+    # via invariant-server
 prometheus-client==0.20.0
     # via prometheus-fastapi-instrumentator
 prometheus-fastapi-instrumentator==7.0.0
     # via invariant-server
 psutil==5.9.8
     # via fastapi-utils
 pydantic==2.8.2
+    # via confection
     # via fastapi
     # via fastapi-utils
     # via invariant
     # via pydantic-settings
+    # via spacy
+    # via thinc
+    # via weasel
 pydantic-core==2.20.1
     # via pydantic
 pydantic-settings==2.3.4
@@ -136,12 +202,26 @@ python-dotenv==1.0.1
 python-multipart==0.0.9
     # via fastapi
 pyyaml==6.0.1
+    # via huggingface-hub
+    # via presidio-analyzer
+    # via transformers
     # via uvicorn
 referencing==0.35.1
     # via jsonschema
     # via jsonschema-specifications
+regex==2024.7.24
+    # via presidio-analyzer
+    # via transformers
 requests==2.32.3
+    # via huggingface-hub
+    # via requests-file
     # via semgrep
+    # via spacy
+    # via tldextract
+    # via transformers
+    # via weasel
+requests-file==2.1.0
+    # via tldextract
 rich==13.7.1
     # via semgrep
     # via typer
@@ -153,28 +233,69 @@ ruamel-yaml==0.17.40
 ruamel-yaml-clib==0.2.8
     # via ruamel-yaml
 ruff==0.5.1
+safetensors==0.4.4
+    # via transformers
 semgrep==1.79.0
     # via invariant
+setuptools==74.1.1
+    # via marisa-trie
+    # via spacy
+    # via thinc
+    # via torch
 shellingham==1.5.4
     # via typer
+smart-open==7.0.4
+    # via weasel
 sniffio==1.3.1
     # via anyio
     # via httpx
+spacy==3.7.6
+    # via invariant-server
+    # via presidio-analyzer
+spacy-legacy==3.0.12
+    # via spacy
+spacy-loggers==1.0.5
+    # via spacy
+srsly==2.4.8
+    # via confection
+    # via spacy
+    # via thinc
+    # via weasel
 starlette==0.37.2
     # via fastapi
     # via prometheus-fastapi-instrumentator
+sympy==1.13.2
+    # via torch
 termcolor==2.4.0
     # via invariant
+thinc==8.2.5
+    # via spacy
+tldextract==5.1.2
+    # via presidio-analyzer
+tokenizers==0.19.1
+    # via transformers
 tomli==2.0.1
     # via semgrep
+torch==2.4.0
+    # via invariant-server
+tqdm==4.66.5
+    # via huggingface-hub
+    # via spacy
+    # via transformers
+transformers==4.44.2
+    # via invariant-server
 typer==0.12.3
     # via fastapi-cli
+    # via spacy
+    # via weasel
 typing-extensions==4.12.2
     # via aiosqlite
     # via fastapi
+    # via huggingface-hub
     # via pydantic
     # via pydantic-core
     # via semgrep
+    # via torch
     # via typer
     # via typing-inspect
 typing-inspect==0.9.0
@@ -188,9 +309,17 @@ uvicorn==0.30.1
     # via invariant-server
 uvloop==0.19.0
     # via uvicorn
+wasabi==1.1.3
+    # via spacy
+    # via thinc
+    # via weasel
 watchfiles==0.22.0
     # via uvicorn
 wcmatch==8.5.2
     # via semgrep
+weasel==0.4.1
+    # via spacy
 websockets==12.0
     # via uvicorn
+wrapt==1.16.0
+    # via smart-open