update

pyproject.toml

@@ -1,8 +1,3 @@
-[[tool.poetry.source]]
-name = "aliyun"
-url = "https://mirrors.aliyun.com/pypi/simple"
-priority = "primary"
-
 [build-system]
 requires = ["poetry-core", "setuptools", "wheel"]
 build-backend = "poetry.core.masonry.api"
@@ -10,12 +5,19 @@ build-backend = "poetry.core.masonry.api"
 [tool.poetry]
 name = "r2r"
 readme = "README.md"
-version = "3.3.14"
+version = "3.3.18"
 
 description = "SciPhi R2R"
 authors = ["Owen Colegrove <owen@sciphi.ai>"]
 license = "MIT"
-include = ["r2r.toml", "compose.yaml", "compose.full.yaml", "pyproject.toml", "migrations/**/*" ]
+include = [
+    "r2r.toml",
+    "compose.yaml",
+    "compose.full.yaml",
+    "compose.full_with_replicas.yaml",
+    "pyproject.toml",
+    "migrations/**/*"
+]
 packages = [
     { include = "r2r" },
     { include = "sdk", from = "." },
@@ -41,7 +43,7 @@ python-dotenv = "^1.0.1"
 requests = "^2.31.0"
 toml = "^0.10.2"
 types-requests = "^2.31.0"
-unstructured-client = "^0.25.5"
+unstructured-client = "0.25.5"
 psycopg-binary = "^3.2.3"
 aiosmtplib = "^3.0.2"
 types-aiofiles = "^24.1.0.20240626"
@@ -82,14 +84,23 @@ aiofiles = { version = "^24.1.0", optional = true }
 aioshutil = { version = "^1.5", optional = true }
 beautifulsoup4 = { version = "^4.12.3", optional = true }
 bs4 = { version = "^0.0.2", optional = true }
-numpy = { version = ">=1.22.4,<1.29.0", optional = true }
+epub = { version = "^0.5.2", optional = true }
+extract-msg = { version = "^0.52.0", optional = true }
 markdown = { version = "^3.6", optional = true }
+numpy = { version = ">=1.22.4,<1.29.0", optional = true }
+olefile = { version = "^0.47", optional = true }
 openpyxl = { version = "^3.1.2", optional = true }
+orgparse = { version = "^0.4.20231004", optional = true }
 pdf2image = { version = "^1.17.0", optional = true }
+pillow-heif = { version = "^0.21.0", optional = true }
 pypdf = { version = "^4.2.0", optional = true }
 pypdf2 = { version = "^3.0.1", optional = true }
 python-pptx = { version = "^1.0.1", optional = true }
 python-docx = { version = "^1.1.0", optional = true }
+docutils = { version = "^0.21.2", optional = true }
+striprtf = { version = "^0.0.28", optional = true }
+xlrd = { version = "^2.0.1", optional = true }
+pynacl = { version = "^1.5.0", optional = true }
 
 [tool.poetry.extras]
 core = [
@@ -121,6 +132,7 @@ core = [
     "uvicorn",
     "vecs",
     "python-multipart",
+    "pynacl"
 ]
 ingestion-bundle = [
     "aiofiles",
@@ -128,13 +140,22 @@ ingestion-bundle = [
     "beautifulsoup4",
     "bs4",
     "numpy",
+    "docutils",
+    "epub",
+    "extract-msg",
     "markdown",
+    "numpy",
+    "olefile",
     "openpyxl",
+    "orgparse",
     "pdf2image",
+    "pillow-heif",
     "pypdf",
     "pypdf2",
     "python-pptx",
     "python-docx",
+    "striprtf",
+    "xlrd",
 ]
 
 [tool.poetry.group.dev.dependencies]
@@ -150,6 +171,7 @@ pytest-mock = "^3.14.0"
 pytest-cov = "^5.0.0"
 pytest-html = "^4.1.1"
 types-toml = "^0.10.8"
+pytest-xdist = "^3.6.1"
 
 [tool.black]
 line-length = 79

r2r.toml

@@ -1,6 +1,9 @@
 [app]
 # app settings are global available like `r2r_config.agent.app`
 # project_name = "r2r_default" # optional, can also set with `R2R_PROJECT_NAME` env var
+default_max_documents_per_user = 100
+default_max_chunks_per_user = 100_000
+default_max_collections_per_user = 10
 
 [agent]
 system_instruction_name = "rag_agent"
@@ -14,10 +17,10 @@ tool_names = ["local_search"]
 provider = "r2r"
 access_token_lifetime_in_minutes = 60
 refresh_token_lifetime_in_days = 7
-require_authentication = true
-require_email_verification = true
-default_admin_email = "xujiawei@cocorobo.cc"
-default_admin_password = "usestudio-1"
+require_authentication = false
+require_email_verification = false
+default_admin_email = "admin@example.com"
+default_admin_password = "change_me_immediately"
 
 [completion]
 provider = "litellm"
@@ -37,8 +40,9 @@ provider = "bcrypt"
 [database]
 provider = "postgres"
 default_collection_name = "Default"
-default_collection_description = "cocorobo collection."
-enable_fts = true # whether or not to enable full-text search, e.g `hybrid search`
+default_collection_description = "Your default collection."
+# collection_summary_system_prompt = 'default_system'
+# collection_summary_task_prompt = 'default_collection_summary'
 
 # KG settings
 batch_size = 256
@@ -87,10 +91,12 @@ provider = "litellm"
 # use `openai/text-embedding-3-large` with `base_dimension = 3072` and binary quantization
 base_model = "openai/text-embedding-3-large"
 base_dimension = 3072
+
 # rerank_model = "huggingface/mixedbread-ai/mxbai-rerank-large-v1" # reranking model
+
 batch_size = 128
 add_title_as_prefix = false
-concurrent_request_limit = 25600
+concurrent_request_limit = 256
 quantization_settings = { quantization_type = "FP32" }
 
 [file]