Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

Refactor: remove API_PROXY_SCHEME (#20470)

Lynnfc9a58caceb4
diff --git a/.github/workflows/sep-tests.yml b/.github/workflows/sep-tests.yml
index 8c681fa2b6..30fb414164 100644
--- a/.github/workflows/sep-tests.yml
+++ b/.github/workflows/sep-tests.yml
@@ -53,7 +53,6 @@ jobs:
has_go_changes: ${{ steps.detect_changes.outputs.has_go_changes }}
has_web_changes: ${{ steps.detect_changes.outputs.has_web_changes }}
has_native_deepdoc_changes: ${{ steps.detect_changes.outputs.has_native_deepdoc_changes }}
- api_proxy_schemes: ${{ steps.detect_changes.outputs.api_proxy_schemes }}
steps:
- name: Ensure workspace ownership
run: |
@@ -213,27 +212,19 @@ jobs:
fi
fi
- api_proxy_schemes='["go"]'
-
echo "has_go_changes=${has_go}" >> $GITHUB_OUTPUT
echo "has_web_changes=${has_web}" >> $GITHUB_OUTPUT
echo "has_native_deepdoc_changes=${has_native_deepdoc}" >> $GITHUB_OUTPUT
- echo "api_proxy_schemes=${api_proxy_schemes}" >> $GITHUB_OUTPUT
- echo "Go: ${has_go}, Web: ${has_web}, Native DeepDoc: ${has_native_deepdoc}, proxy schemes: ${api_proxy_schemes}"
+ echo "Go: ${has_go}, Web: ${has_web}, Native DeepDoc: ${has_native_deepdoc}"
ragflow_tests_infinity:
- name: ragflow_tests_infinity (${{ matrix.api_proxy_scheme }})
+ name: ragflow_tests_infinity
needs: ragflow_preflight
if: ${{ github.event_name != 'pull_request' && github.event_name != 'pull_request_target' || (github.event.pull_request.draft == false && contains(github.event.pull_request.labels.*.name, 'ci') && (github.event.action != 'labeled' || github.event.label.name == 'ci')) }}
runs-on: [ "self-hosted", "ragflow-test" ]
- strategy:
- fail-fast: false
- matrix:
- api_proxy_scheme: ${{ fromJSON(needs.ragflow_preflight.outputs.api_proxy_schemes) }}
env:
DOC_ENGINE: infinity
- API_PROXY_SCHEME: ${{ matrix.api_proxy_scheme }}
- RAGFLOW_IMAGE: infiniflow/ragflow:${{ github.run_id }}-infinity-${{ matrix.api_proxy_scheme }}
+ RAGFLOW_IMAGE: infiniflow/ragflow:${{ github.run_id }}-infinity
HTTP_API_TEST_LEVEL: ${{ needs.ragflow_preflight.outputs.http_api_test_level }}
HAS_GO: ${{ needs.ragflow_preflight.outputs.has_go_changes }}
HAS_WEB: ${{ needs.ragflow_preflight.outputs.has_web_changes }}
@@ -254,7 +245,6 @@ jobs:
allow-unsafe-pr-checkout: true
- name: Build ragflow go server
- if: env.API_PROXY_SCHEME == 'go'
run: |
set -euo pipefail
BUILDER_CONTAINER=ragflow_build_${GITHUB_RUN_ID}_${DOC_ENGINE}_$(od -An -N4 -tx4 /dev/urandom | tr -d ' ')
@@ -276,7 +266,6 @@ jobs:
./build.sh --go
- name: Run Go unit tests
- if: env.API_PROXY_SCHEME == 'go'
# Runs after `./build.sh --go`, which guarantees the C++ static
# library (librag_tokenizer_c_api.a) is present on disk. The Go
# test binaries link against it transitively through
@@ -388,19 +377,14 @@ jobs:
# Dockerfile / Dockerfile_ci read it from ARG GITEE_TOKEN (NOT a BuildKit secret),
# so the runner must expose GITEE_TOKEN as an environment variable (e.g. via the
# runner's docker-compose `environment:` mapping sourced from its .env file).
- # Pick the Dockerfile matching the API proxy scheme:
- # - go -> Dockerfile (Go-native image)
- # - * -> Dockerfile_ci (Python/legacy image)
- DOCKERFILE=Dockerfile
(
flock -w 10800 9 || { echo "Timed out waiting for the shared Docker build slot" >&2; exit 1; }
- echo "Acquired Docker build slot for ${DOC_ENGINE}/${API_PROXY_SCHEME} (dockerfile=${DOCKERFILE})"
+ echo "Acquired Docker build slot for ${DOC_ENGINE}"
##sudo docker pull ubuntu:24.04
- sudo DOCKER_BUILDKIT=1 docker build --build-arg NEED_MIRROR=1 --build-arg HTTPS_PROXY=${HTTPS_PROXY} --build-arg HTTP_PROXY=${HTTP_PROXY} --build-arg GITEE_TOKEN=${GITEE_TOKEN} -f "${DOCKERFILE}" -t ${RAGFLOW_IMAGE} .
+ sudo DOCKER_BUILDKIT=1 docker build --build-arg NEED_MIRROR=1 --build-arg HTTPS_PROXY=${HTTPS_PROXY} --build-arg HTTP_PROXY=${HTTP_PROXY} --build-arg GITEE_TOKEN=${GITEE_TOKEN} -f Dockerfile -t ${RAGFLOW_IMAGE} .
) 9>"${BUILD_LOCK_FILE}"
- name: Prepare Python test environment
- if: env.API_PROXY_SCHEME != 'web'
run: |
git config --global url."https://oauth2:${GITEE_TOKEN}@gitee.com/".insteadOf "https://gitee.com/"
uv sync --python 3.13 --group test --frozen
@@ -414,10 +398,10 @@ jobs:
## ss already installed
##sudo apt update && sudo apt install -y iproute2
RUNNER_WORKSPACE_PREFIX=${RUNNER_WORKSPACE_PREFIX:-${HOME}}
- COMPOSE_PROJECT_NAME="${GITHUB_RUN_ID}-${DOC_ENGINE}-${API_PROXY_SCHEME}"
+ COMPOSE_PROJECT_NAME="${GITHUB_RUN_ID}-${DOC_ENGINE}"
echo "COMPOSE_PROJECT_NAME=${COMPOSE_PROJECT_NAME}" >> ${GITHUB_ENV}
echo "RAGFLOW_CONTAINER=${COMPOSE_PROJECT_NAME}-ragflow-cpu-1" >> ${GITHUB_ENV}
- ARTIFACTS_DIR=${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}/${API_PROXY_SCHEME}
+ ARTIFACTS_DIR=${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}
echo "ARTIFACTS_DIR=${ARTIFACTS_DIR}" >> ${GITHUB_ENV}
rm -rf "${ARTIFACTS_DIR}" && mkdir -p "${ARTIFACTS_DIR}"
@@ -607,7 +591,7 @@ jobs:
# Persist computed ports into .env so docker-compose uses the correct host bindings.
# Remove previous CI overrides first; docker compose uses the last duplicate key.
- sed -i '/^ES_PORT=/d;/^OS_PORT=/d;/^INFINITY_THRIFT_PORT=/d;/^INFINITY_HTTP_PORT=/d;/^INFINITY_PSQL_PORT=/d;/^EXPOSE_MYSQL_PORT=/d;/^MINIO_PORT=/d;/^MINIO_CONSOLE_PORT=/d;/^REDIS_PORT=/d;/^KVROCKS_PORT=/d;/^EXPOSE_NATS_PORT=/d;/^EXPOSE_NATS_MONITORING_PORT=/d;/^NATS_MONITORING_PORT=/d;/^TEI_PORT=/d;/^KIBANA_PORT=/d;/^SVR_HTTP_PORT=/d;/^ADMIN_SVR_HTTP_PORT=/d;/^SVR_MCP_PORT=/d;/^SANDBOX_EXECUTOR_MANAGER_PORT=/d;/^SVR_WEB_HTTP_PORT=/d;/^SVR_WEB_HTTPS_PORT=/d;/^EXPOSE_CLICKHOUSE_TCP_PORT=/d;/^CLICKHOUSE_HTTP_PORT=/d;/^COMPOSE_PROFILES=/d;/^TEI_MODEL=/d;/^RAGFLOW_IMAGE=/d;/^DOC_ENGINE=/d;/^API_PROXY_SCHEME=/d' .env
+ sed -i '/^ES_PORT=/d;/^OS_PORT=/d;/^INFINITY_THRIFT_PORT=/d;/^INFINITY_HTTP_PORT=/d;/^INFINITY_PSQL_PORT=/d;/^EXPOSE_MYSQL_PORT=/d;/^MINIO_PORT=/d;/^MINIO_CONSOLE_PORT=/d;/^REDIS_PORT=/d;/^KVROCKS_PORT=/d;/^EXPOSE_NATS_PORT=/d;/^EXPOSE_NATS_MONITORING_PORT=/d;/^NATS_MONITORING_PORT=/d;/^TEI_PORT=/d;/^KIBANA_PORT=/d;/^SVR_HTTP_PORT=/d;/^ADMIN_SVR_HTTP_PORT=/d;/^SVR_MCP_PORT=/d;/^SANDBOX_EXECUTOR_MANAGER_PORT=/d;/^SVR_WEB_HTTP_PORT=/d;/^SVR_WEB_HTTPS_PORT=/d;/^EXPOSE_CLICKHOUSE_TCP_PORT=/d;/^CLICKHOUSE_HTTP_PORT=/d;/^COMPOSE_PROFILES=/d;/^TEI_MODEL=/d;/^RAGFLOW_IMAGE=/d;/^DOC_ENGINE=/d' .env
# docker-compose-go.yml uses env_file: .env-go for the ragflow-cpu
# container, so the Go server process reads COMPOSE_PROFILES and
# TEI_MODEL from .env (not .env). Without syncing these two,
@@ -644,7 +628,6 @@ jobs:
echo "TEI_MODEL=BAAI/bge-small-en-v1.5"
echo "RAGFLOW_IMAGE=${RAGFLOW_IMAGE}"
echo "DOC_ENGINE=${DOC_ENGINE}"
- echo "API_PROXY_SCHEME=${API_PROXY_SCHEME}"
} >> .env
# Sync COMPOSE_PROFILES and TEI_MODEL into .env so the Go
# server container (env_file: .env) sees the same TEI bypass
@@ -675,7 +658,6 @@ jobs:
sudo docker compose -f "${COMPOSE_FILE}" -p ${COMPOSE_PROJECT_NAME} up -d
- name: Run New RESTFUL api tests against Infinity
- if: env.API_PROXY_SCHEME != 'web'
run: |
export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
svc_ready=0
@@ -710,7 +692,7 @@ jobs:
run: |
set -uo pipefail
RUNNER_WORKSPACE_PREFIX="${RUNNER_WORKSPACE_PREFIX:-${HOME}}"
- ARTIFACTS_DIR="${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}/${API_PROXY_SCHEME}"
+ ARTIFACTS_DIR="${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}"
LOG_DEST="${ARTIFACTS_DIR}/ragflow-logs-infinity"
mkdir -p "${ARTIFACTS_DIR}"
if [ -d docker/ragflow-logs ]; then
@@ -749,18 +731,13 @@ jobs:
ragflow_tests_elasticsearch:
- name: ragflow_tests_elasticsearch (${{ matrix.api_proxy_scheme }})
+ name: ragflow_tests_elasticsearch
needs: ragflow_preflight
if: ${{ github.event_name != 'pull_request' && github.event_name != 'pull_request_target' || (github.event.pull_request.draft == false && contains(github.event.pull_request.labels.*.name, 'ci') && (github.event.action != 'labeled' || github.event.label.name == 'ci')) }}
runs-on: [ "self-hosted", "ragflow-test" ]
- strategy:
- fail-fast: false
- matrix:
- api_proxy_scheme: ${{ fromJSON(needs.ragflow_preflight.outputs.api_proxy_schemes) }}
env:
DOC_ENGINE: elasticsearch
- API_PROXY_SCHEME: ${{ matrix.api_proxy_scheme }}
- RAGFLOW_IMAGE: infiniflow/ragflow:${{ github.run_id }}-elasticsearch-${{ matrix.api_proxy_scheme }}
+ RAGFLOW_IMAGE: infiniflow/ragflow:${{ github.run_id }}-elasticsearch
HTTP_API_TEST_LEVEL: ${{ needs.ragflow_preflight.outputs.http_api_test_level }}
HAS_GO: ${{ needs.ragflow_preflight.outputs.has_go_changes }}
HAS_WEB: ${{ needs.ragflow_preflight.outputs.has_web_changes }}
@@ -781,7 +758,6 @@ jobs:
allow-unsafe-pr-checkout: true
- name: Build ragflow go server
- if: env.API_PROXY_SCHEME == 'go'
run: |
set -euo pipefail
BUILDER_CONTAINER=ragflow_build_${GITHUB_RUN_ID}_${DOC_ENGINE}_$(od -An -N4 -tx4 /dev/urandom | tr -d ' ')
@@ -803,7 +779,6 @@ jobs:
./build.sh --go
- name: Run Go unit tests
- if: env.API_PROXY_SCHEME == 'go'
# Runs after `./build.sh --go`, which guarantees the C++ static
# library (librag_tokenizer_c_api.a) is present on disk. The Go
# test binaries link against it transitively through
@@ -906,7 +881,6 @@ jobs:
fi
- name: Run ragflow_deps regression guards
- if: env.API_PROXY_SCHEME == 'go'
# Pure-Python, local, needs no API keys or model downloads; guards (a) the
# ONNX Runtime extraction logic the in-process (Go) DeepDoc backend build
# depends on (notably the version-bump path C2 hardened) and (b) the
@@ -930,19 +904,14 @@ jobs:
# Dockerfile / Dockerfile_ci read it from ARG GITEE_TOKEN (NOT a BuildKit secret),
# so the runner must expose GITEE_TOKEN as an environment variable (e.g. via the
# runner's docker-compose `environment:` mapping sourced from its .env file).
- # Pick the Dockerfile matching the API proxy scheme:
- # - go -> Dockerfile (Go-native image)
- # - * -> Dockerfile_ci (Python/legacy image)
- DOCKERFILE=Dockerfile
(
flock -w 10800 9 || { echo "Timed out waiting for the shared Docker build slot" >&2; exit 1; }
- echo "Acquired Docker build slot for ${DOC_ENGINE}/${API_PROXY_SCHEME} (dockerfile=${DOCKERFILE})"
+ echo "Acquired Docker build slot for ${DOC_ENGINE}"
##sudo docker pull ubuntu:24.04
- sudo DOCKER_BUILDKIT=1 docker build --build-arg NEED_MIRROR=1 --build-arg HTTPS_PROXY=${HTTPS_PROXY} --build-arg HTTP_PROXY=${HTTP_PROXY} --build-arg GITEE_TOKEN=${GITEE_TOKEN} -f "${DOCKERFILE}" -t ${RAGFLOW_IMAGE} .
+ sudo DOCKER_BUILDKIT=1 docker build --build-arg NEED_MIRROR=1 --build-arg HTTPS_PROXY=${HTTPS_PROXY} --build-arg HTTP_PROXY=${HTTP_PROXY} --build-arg GITEE_TOKEN=${GITEE_TOKEN} -f Dockerfile -t ${RAGFLOW_IMAGE} .
) 9>"${BUILD_LOCK_FILE}"
- name: Prepare Python test environment
- if: env.API_PROXY_SCHEME != 'web'
run: |
git config --global url."https://oauth2:${GITEE_TOKEN}@gitee.com/".insteadOf "https://gitee.com/"
uv sync --python 3.13 --group test --frozen
@@ -956,10 +925,10 @@ jobs:
## ss already installed
##sudo apt update && sudo apt install -y iproute2
RUNNER_WORKSPACE_PREFIX=${RUNNER_WORKSPACE_PREFIX:-${HOME}}
- COMPOSE_PROJECT_NAME="${GITHUB_RUN_ID}-${DOC_ENGINE}-${API_PROXY_SCHEME}"
+ COMPOSE_PROJECT_NAME="${GITHUB_RUN_ID}-${DOC_ENGINE}"
echo "COMPOSE_PROJECT_NAME=${COMPOSE_PROJECT_NAME}" >> ${GITHUB_ENV}
echo "RAGFLOW_CONTAINER=${COMPOSE_PROJECT_NAME}-ragflow-cpu-1" >> ${GITHUB_ENV}
- ARTIFACTS_DIR=${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}/${API_PROXY_SCHEME}
+ ARTIFACTS_DIR=${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}
echo "ARTIFACTS_DIR=${ARTIFACTS_DIR}" >> ${GITHUB_ENV}
rm -rf "${ARTIFACTS_DIR}" && mkdir -p "${ARTIFACTS_DIR}"
@@ -1149,7 +1118,7 @@ jobs:
# Persist computed ports into .env so docker-compose uses the correct host bindings.
# Remove previous CI overrides first; docker compose uses the last duplicate key.
- sed -i '/^ES_PORT=/d;/^OS_PORT=/d;/^INFINITY_THRIFT_PORT=/d;/^INFINITY_HTTP_PORT=/d;/^INFINITY_PSQL_PORT=/d;/^EXPOSE_MYSQL_PORT=/d;/^MINIO_PORT=/d;/^MINIO_CONSOLE_PORT=/d;/^REDIS_PORT=/d;/^KVROCKS_PORT=/d;/^EXPOSE_NATS_PORT=/d;/^EXPOSE_NATS_MONITORING_PORT=/d;/^NATS_MONITORING_PORT=/d;/^TEI_PORT=/d;/^KIBANA_PORT=/d;/^SVR_HTTP_PORT=/d;/^ADMIN_SVR_HTTP_PORT=/d;/^SVR_MCP_PORT=/d;/^SANDBOX_EXECUTOR_MANAGER_PORT=/d;/^SVR_WEB_HTTP_PORT=/d;/^SVR_WEB_HTTPS_PORT=/d;/^EXPOSE_CLICKHOUSE_TCP_PORT=/d;/^CLICKHOUSE_HTTP_PORT=/d;/^COMPOSE_PROFILES=/d;/^TEI_MODEL=/d;/^RAGFLOW_IMAGE=/d;/^DOC_ENGINE=/d;/^API_PROXY_SCHEME=/d' .env
+ sed -i '/^ES_PORT=/d;/^OS_PORT=/d;/^INFINITY_THRIFT_PORT=/d;/^INFINITY_HTTP_PORT=/d;/^INFINITY_PSQL_PORT=/d;/^EXPOSE_MYSQL_PORT=/d;/^MINIO_PORT=/d;/^MINIO_CONSOLE_PORT=/d;/^REDIS_PORT=/d;/^KVROCKS_PORT=/d;/^EXPOSE_NATS_PORT=/d;/^EXPOSE_NATS_MONITORING_PORT=/d;/^NATS_MONITORING_PORT=/d;/^TEI_PORT=/d;/^KIBANA_PORT=/d;/^SVR_HTTP_PORT=/d;/^ADMIN_SVR_HTTP_PORT=/d;/^SVR_MCP_PORT=/d;/^SANDBOX_EXECUTOR_MANAGER_PORT=/d;/^SVR_WEB_HTTP_PORT=/d;/^SVR_WEB_HTTPS_PORT=/d;/^EXPOSE_CLICKHOUSE_TCP_PORT=/d;/^CLICKHOUSE_HTTP_PORT=/d;/^COMPOSE_PROFILES=/d;/^TEI_MODEL=/d;/^RAGFLOW_IMAGE=/d;/^DOC_ENGINE=/d' .env
# docker-compose-go.yml uses env_file: .env-go for the ragflow-cpu
# container, so the Go server process reads COMPOSE_PROFILES and
# TEI_MODEL from .env (not .env). Without syncing these two,
@@ -1186,7 +1155,6 @@ jobs:
echo "TEI_MODEL=BAAI/bge-small-en-v1.5"
echo "RAGFLOW_IMAGE=${RAGFLOW_IMAGE}"
echo "DOC_ENGINE=${DOC_ENGINE}"
- echo "API_PROXY_SCHEME=${API_PROXY_SCHEME}"
} >> .env
# Sync COMPOSE_PROFILES and TEI_MODEL into .env so the Go
# server container (env_file: .env) sees the same TEI bypass
@@ -1216,7 +1184,6 @@ jobs:
sudo docker compose -f "${COMPOSE_FILE}" -p ${COMPOSE_PROJECT_NAME} up -d
- name: Run New RESTFUL api tests against Elasticsearch
- if: env.API_PROXY_SCHEME != 'web'
run: |
export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
svc_ready=0
@@ -1243,7 +1210,7 @@ jobs:
run: |
set -uo pipefail
RUNNER_WORKSPACE_PREFIX="${RUNNER_WORKSPACE_PREFIX:-${HOME}}"
- ARTIFACTS_DIR="${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}/${API_PROXY_SCHEME}"
+ ARTIFACTS_DIR="${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}"
LOG_DEST="${ARTIFACTS_DIR}/ragflow-logs-es"
mkdir -p "${ARTIFACTS_DIR}"
if [ -d docker/ragflow-logs ]; then
@@ -1390,7 +1357,6 @@ jobs:
runs-on: [ "self-hosted", "ragflow-test" ]
env:
DOC_ENGINE: infinity
- API_PROXY_SCHEME: go
steps:
- name: Ensure workspace ownership
run: |
@@ -1436,10 +1402,10 @@ jobs:
## ss already installed
##sudo apt update && sudo apt install -y iproute2
RUNNER_WORKSPACE_PREFIX=${RUNNER_WORKSPACE_PREFIX:-${HOME}}
- COMPOSE_PROJECT_NAME="${GITHUB_RUN_ID}-go-int-${DOC_ENGINE}-${API_PROXY_SCHEME}"
+ COMPOSE_PROJECT_NAME="${GITHUB_RUN_ID}-go-int-${DOC_ENGINE}"
echo "COMPOSE_PROJECT_NAME=${COMPOSE_PROJECT_NAME}" >> ${GITHUB_ENV}
echo "RAGFLOW_CONTAINER=${COMPOSE_PROJECT_NAME}-ragflow-cpu-1" >> ${GITHUB_ENV}
- ARTIFACTS_DIR=${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}/${API_PROXY_SCHEME}
+ ARTIFACTS_DIR=${RUNNER_WORKSPACE_PREFIX}/artifacts/${GITHUB_REPOSITORY}/${GITHUB_RUN_ID}/${DOC_ENGINE}
echo "ARTIFACTS_DIR=${ARTIFACTS_DIR}" >> ${GITHUB_ENV}
rm -rf "${ARTIFACTS_DIR}" && mkdir -p "${ARTIFACTS_DIR}"
@@ -1629,7 +1595,7 @@ jobs:
# Persist computed ports into .env so docker-compose uses the correct host bindings.
# Remove previous CI overrides first; docker compose uses the last duplicate key.
- sed -i '/^ES_PORT=/d;/^OS_PORT=/d;/^INFINITY_THRIFT_PORT=/d;/^INFINITY_HTTP_PORT=/d;/^INFINITY_PSQL_PORT=/d;/^EXPOSE_MYSQL_PORT=/d;/^MINIO_PORT=/d;/^MINIO_CONSOLE_PORT=/d;/^REDIS_PORT=/d;/^KVROCKS_PORT=/d;/^EXPOSE_NATS_PORT=/d;/^EXPOSE_NATS_MONITORING_PORT=/d;/^NATS_MONITORING_PORT=/d;/^TEI_PORT=/d;/^KIBANA_PORT=/d;/^SVR_HTTP_PORT=/d;/^ADMIN_SVR_HTTP_PORT=/d;/^SVR_MCP_PORT=/d;/^SANDBOX_EXECUTOR_MANAGER_PORT=/d;/^SVR_WEB_HTTP_PORT=/d;/^SVR_WEB_HTTPS_PORT=/d;/^EXPOSE_CLICKHOUSE_TCP_PORT=/d;/^CLICKHOUSE_HTTP_PORT=/d;/^COMPOSE_PROFILES=/d;/^TEI_MODEL=/d;/^RAGFLOW_IMAGE=/d;/^DOC_ENGINE=/d;/^API_PROXY_SCHEME=/d' .env
+ sed -i '/^ES_PORT=/d;/^OS_PORT=/d;/^INFINITY_THRIFT_PORT=/d;/^INFINITY_HTTP_PORT=/d;/^INFINITY_PSQL_PORT=/d;/^EXPOSE_MYSQL_PORT=/d;/^MINIO_PORT=/d;/^MINIO_CONSOLE_PORT=/d;/^REDIS_PORT=/d;/^KVROCKS_PORT=/d;/^EXPOSE_NATS_PORT=/d;/^EXPOSE_NATS_MONITORING_PORT=/d;/^NATS_MONITORING_PORT=/d;/^TEI_PORT=/d;/^KIBANA_PORT=/d;/^SVR_HTTP_PORT=/d;/^ADMIN_SVR_HTTP_PORT=/d;/^SVR_MCP_PORT=/d;/^SANDBOX_EXECUTOR_MANAGER_PORT=/d;/^SVR_WEB_HTTP_PORT=/d;/^SVR_WEB_HTTPS_PORT=/d;/^EXPOSE_CLICKHOUSE_TCP_PORT=/d;/^CLICKHOUSE_HTTP_PORT=/d;/^COMPOSE_PROFILES=/d;/^TEI_MODEL=/d;/^RAGFLOW_IMAGE=/d;/^DOC_ENGINE=/d' .env
# docker-compose-go.yml uses env_file: .env-go for the ragflow-cpu
# container, so the Go server process reads COMPOSE_PROFILES and
# TEI_MODEL from .env (not .env). Without syncing these two,
@@ -1672,7 +1638,6 @@ jobs:
echo "COMPOSE_PROFILES=infinity,ragflow-go,tei-cpu,metadata-mysql,clickhouse"
echo "TEI_MODEL=BAAI/bge-small-en-v1.5"
echo "DOC_ENGINE=${DOC_ENGINE}"
- echo "API_PROXY_SCHEME=${API_PROXY_SCHEME}"
} >> .env
# Sync COMPOSE_PROFILES and TEI_MODEL into .env so the Go
# server container (env_file: .env) sees the same TEI bypass
diff --git a/docker/.env b/docker/.env
index 935730d119..d0d16190eb 100644
--- a/docker/.env
+++ b/docker/.env
@@ -264,10 +264,6 @@ SVR_MCP_PORT=9382
# first login.
ADMIN_DEFAULT_PASSWORD=admin
-# API_PROXY_SCHEME=hybrid # go and python hybrid deploy mode
-API_PROXY_SCHEME=go
-#API_PROXY_SCHEME=python # use pure python server deployment
-
# Development-only: set to 1 to bypass host safety checks for test_db_connection and allow private/local database hosts.
# Do not enable in production.
ALLOW_ANY_HOST=0
diff --git a/docker/launch_backend_service.sh b/docker/launch_backend_service.sh
index eba2d5f96e..3ab864ad2f 100755
--- a/docker/launch_backend_service.sh
+++ b/docker/launch_backend_service.sh
@@ -12,10 +12,10 @@ usage() {
echo " --debug Enable debug mode and debug-level logging"
echo
echo "Available service types:"
- echo " ragflow Start RAGFlow server based on API_PROXY_SCHEME"
- echo " task_executor Start task workers based on API_PROXY_SCHEME"
- echo " admin Start Admin server based on API_PROXY_SCHEME"
- echo " data_sync Start the data source syncer based on API_PROXY_SCHEME"
+ echo " ragflow Start RAGFlow server"
+ echo " task_executor Start task workers"
+ echo " admin Start Admin server"
+ echo " data_sync Start the data source syncer"
echo
echo "Examples:"
echo " $0"
@@ -50,12 +50,8 @@ load_env_file
# Unset HTTP proxies that might be set by Docker daemon
export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
-export PYTHONPATH=$(pwd)
export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu/
-JEMALLOC_PATH=$(pkg-config --variable=libdir jemalloc)/libjemalloc.so
-
-PY=python3
# Set default number of workers if WS is not set or less than 1
if [[ -z "$WS" || $WS -lt 1 ]]; then
@@ -95,25 +91,18 @@ trap cleanup SIGINT SIGTERM
# Function to execute task_executor with retry logic
task_exe(){
local task_id=$1
- local task_name="task_executor.py for task $task_id"
- local -a task_cmd=("$PY" "rag/svr/task_executor.py" "-i" "$task_id")
- if [[ "${API_PROXY_SCHEME}" == "go" ]]; then
- prepare_for_go
- task_name="ragflow_server --ingestor"
- task_cmd=("bin/ragflow_server" "--ingestor")
- if [[ "$DEBUG_MODE" -eq 1 ]]; then
- task_cmd+=("--log-level" "debug")
- fi
+ local task_name="ragflow_server --ingestor"
+ local -a task_cmd=("bin/ragflow_server" "--ingestor")
+ prepare_for_go
+ if [[ "$DEBUG_MODE" -eq 1 ]]; then
+ task_cmd+=("--log-level" "debug")
fi
local retry_count=0
+
while ! $STOP && [ $retry_count -lt $MAX_RETRIES ]; do
echo "Starting ${task_cmd[*]} (Attempt $((retry_count+1)))"
EXIT_CODE=0
- if [[ "${API_PROXY_SCHEME}" == "go" ]]; then
- "${task_cmd[@]}" || EXIT_CODE=$?
- else
- LD_PRELOAD=$JEMALLOC_PATH "${task_cmd[@]}" || EXIT_CODE=$?
- fi
+ "${task_cmd[@]}" || EXIT_CODE=$?
if [ $EXIT_CODE -eq 0 ]; then
echo "$task_name exited successfully."
break
@@ -132,17 +121,11 @@ task_exe(){
# Function to execute ragflow_server with retry logic
run_server(){
- local server_name="ragflow_server.py"
- local -a server_cmd=("$PY" "api/ragflow_server.py")
- if [[ "${API_PROXY_SCHEME}" == "go" ]]; then
- prepare_for_go
- server_name="ragflow_server"
- server_cmd=("bin/ragflow_server" "--api")
- if [[ "$DEBUG_MODE" -eq 1 ]]; then
- server_cmd+=("--log-level" "debug")
- fi
- elif [[ "$DEBUG_MODE" -eq 1 ]]; then
- server_cmd+=("--debug")
+ local server_name="ragflow_server"
+ local -a server_cmd=("bin/ragflow_server" "--api")
+ prepare_for_go
+ if [[ "$DEBUG_MODE" -eq 1 ]]; then
+ server_cmd+=("--log-level" "debug")
fi
local retry_count=0
while ! $STOP && [ $retry_count -lt $MAX_RETRIES ]; do
@@ -167,15 +150,11 @@ run_server(){
# Function to execute admin_server with retry logic
run_admin_server(){
- local server_name="admin_server.py"
- local -a server_cmd=("$PY" "admin/server/admin_server.py")
- if [[ "${API_PROXY_SCHEME}" == "go" ]]; then
- prepare_for_go
- server_name="admin_server"
- server_cmd=("bin/ragflow_server" "--admin")
- if [[ "$DEBUG_MODE" -eq 1 ]]; then
- server_cmd+=("--log-level" "debug")
- fi
+ local server_name="admin_server"
+ local -a server_cmd=("bin/ragflow_server" "--admin")
+ prepare_for_go
+ if [[ "$DEBUG_MODE" -eq 1 ]]; then
+ server_cmd+=("--log-level" "debug")
fi
local retry_count=0
while ! $STOP && [ $retry_count -lt $MAX_RETRIES ]; do
@@ -199,15 +178,11 @@ run_admin_server(){
# Function to execute sync_data_source with retry logic
run_data_sync(){
- local server_name="sync_data_source.py"
- local -a sync_cmd=("$PY" "rag/svr/sync_data_source.py")
- if [[ "${API_PROXY_SCHEME}" == "go" ]]; then
- prepare_for_go
- server_name="ragflow_server --syncer"
- sync_cmd=("bin/ragflow_server" "--syncer")
- if [[ "$DEBUG_MODE" -eq 1 ]]; then
- sync_cmd+=("--log-level" "debug")
- fi
+ local server_name="ragflow_server --syncer"
+ local -a sync_cmd=("bin/ragflow_server" "--syncer")
+ prepare_for_go
+ if [[ "$DEBUG_MODE" -eq 1 ]]; then
+ sync_cmd+=("--log-level" "debug")
fi
local retry_count=0
while ! $STOP && [ $retry_count -lt $MAX_RETRIES ]; do
@@ -230,12 +205,6 @@ run_data_sync(){
fi
}
-ensure_db_init() {
- echo "Initializing database tables..."
- "$PY" -c "from api.db.db_models import init_database_tables as init_web_db; init_web_db()"
- echo "Database tables initialized."
-}
-
run_migrations() {
local db_type="${DB_TYPE:-mysql}"
db_type="${db_type,,}"
@@ -246,15 +215,10 @@ run_migrations() {
return 0
fi
- if [[ "${API_PROXY_SCHEME}" == "go" ]]; then
- # The Go backend owns the model provider tables. --migrate is a
- # standalone action: it runs the migrations and exits, so it is
- # independent of any server actually starting.
- echo "Running model provider table migrations (go)..."
- bin/ragflow_server --migrate
- else
- tools/scripts/run_migrations.sh
- fi
+ # --migrate is a standalone action: it runs the migrations and exits, so it
+ # is independent of any server actually starting.
+ echo "Running model provider table migrations..."
+ bin/ragflow_server --migrate
}
prepare_for_go() {
@@ -317,23 +281,12 @@ for arg in "$@"; do
esac
done
-if [[ "$DEBUG_MODE" -eq 1 && "${API_PROXY_SCHEME}" != "go" ]]; then
- export LOG_LEVELS="root=DEBUG"
-fi
-
if [[ "$SERVICE_SELECTED" -eq 0 ]]; then
START_RAGFLOW=1
START_TASK_EXECUTOR=1
fi
if [[ "$START_RAGFLOW" -eq 1 ]]; then
- # The Go backend owns the schema when it serves the API, so skip the
- # Python-side table creation then. A Go-only deployment may not ship the api
- # package at all, and letting both sides create tables makes them fight over
- # the same schema.
- if [[ "${API_PROXY_SCHEME}" != "go" ]]; then
- ensure_db_init
- fi
run_migrations
fi
diff --git a/internal/handler/dataset.go b/internal/handler/dataset.go
index 7591c6528e..679260743a 100644
--- a/internal/handler/dataset.go
+++ b/internal/handler/dataset.go
@@ -948,9 +948,8 @@ func (h *DatasetsHandler) AggregateTags(c *gin.Context) {
}
// GetCompilationStatus returns the dataset-level knowledge-compile lifecycle
-// state (scheduler contract for API_PROXY_SCHEME=go/hybrid). It replaces the
-// Python-era TraceIndex task-progress endpoint for the Go backend. The optional
-// `kind` query parameter scopes the status to one compile type.
+// state. The optional `kind` query parameter scopes the status to one compile
+// type.
func (h *DatasetsHandler) GetCompilationStatus(c *gin.Context) {
user, errorCode, errorMessage := GetUser(c)
if errorCode != common.CodeSuccess {
diff --git a/internal/router/router.go b/internal/router/router.go
index cf5ed59652..3f38e8529d 100644
--- a/internal/router/router.go
+++ b/internal/router/router.go
@@ -340,8 +340,8 @@ func (r *Router) Setup(engine *gin.Engine) {
datasets.GET("/:dataset_id/graph", r.datasetsHandler.GetKnowledgeGraph)
datasets.POST("/:dataset_id/embedding/check", r.datasetsHandler.CheckEmbedding)
datasets.POST("/:dataset_id/documents/batch-update-status", r.documentHandler.BatchUpdateDocumentStatus)
- // Scheduler compile-status contract (API_PROXY_SCHEME=go/hybrid);
- // replaces the retired RunIndex/TraceIndex/DeleteIndex /index routes.
+ // Scheduler compile-status contract; replaces the retired
+ // RunIndex/TraceIndex/DeleteIndex /index routes.
datasets.GET("/:dataset_id/compilation/status", r.datasetsHandler.GetCompilationStatus)
// Knowledge-compilation wiki artifacts
diff --git a/internal/service/dataset/compilation_status.go b/internal/service/dataset/compilation_status.go
index 6ce2a8f712..9224bde74d 100644
--- a/internal/service/dataset/compilation_status.go
+++ b/internal/service/dataset/compilation_status.go
@@ -15,9 +15,7 @@ import (
)
// CompilationStatus is the dataset-level knowledge-compile lifecycle state
-// surfaced by GET /datasets/:id/compilation/status. It is the Go scheduler
-// contract that replaces the Python-era RunIndex/TraceIndex task progress for
-// API_PROXY_SCHEME=go / hybrid.
+// surfaced by GET /datasets/:id/compilation/status.
//
// State only takes one of idle/pending/running/completed. Error is NOT a fifth
// state: it is a diagnostic attached to a pending/running batch left for retry,
diff --git a/test/testcases/configs.py b/test/testcases/configs.py
index 65b71ef5b9..7d0b8daba2 100644
--- a/test/testcases/configs.py
+++ b/test/testcases/configs.py
@@ -13,43 +13,16 @@
# See the License for the specific language governing permissions and
# limitations under the License.
#
-import logging
import os
-from pathlib import Path
import pytest
-_DOCKER_ENV = Path(__file__).resolve().parents[2] / "docker" / ".env"
-
-
-def _docker_env_value(name: str) -> str | None:
- if not _DOCKER_ENV.exists():
- return None
- for line in _DOCKER_ENV.read_text().splitlines():
- line = line.strip()
- if not line or line.startswith("#") or "=" not in line:
- continue
- key, value = line.split("=", 1)
- if key.strip() == name:
- return value.split("#", 1)[0].strip().strip('"').strip("'")
- return None
-
-
-def _config_value(name: str, default: str | None = None) -> str | None:
- return os.getenv(name) or _docker_env_value(name) or default
-
-
-API_PROXY_SCHEME = _config_value("API_PROXY_SCHEME", "python")
-IS_GO_PROXY = API_PROXY_SCHEME == "go"
-SDK_UNAUTHORIZED_ERROR_MESSAGE = "Invalid access token" if IS_GO_PROXY else "<Unauthorized '401: Unauthorized'>"
-
def _default_host_address() -> str:
return "http://127.0.0.1:9380"
HOST_ADDRESS = os.getenv("HOST_ADDRESS") or _default_host_address()
-logging.info("Resolved API proxy configuration: scheme=%s, is_go_proxy=%s, host_address=%s", API_PROXY_SCHEME, IS_GO_PROXY, HOST_ADDRESS)
VERSION = "v1"
ZHIPU_AI_API_KEY = os.getenv("ZHIPU_AI_API_KEY")
if ZHIPU_AI_API_KEY is None:
diff --git a/test/testcases/conftest.py b/test/testcases/conftest.py
index 09439d96e5..dc14d09a8b 100644
--- a/test/testcases/conftest.py
+++ b/test/testcases/conftest.py
@@ -95,7 +95,7 @@ import time
import pytest
import requests
-from test.testcases.configs import API_PROXY_SCHEME, EMAIL, HOST_ADDRESS, IS_GO_PROXY, PASSWORD, SILICONFLOW_API_KEY, VERSION, ZHIPU_AI_API_KEY
+from test.testcases.configs import EMAIL, HOST_ADDRESS, PASSWORD, SILICONFLOW_API_KEY, VERSION, ZHIPU_AI_API_KEY
MARKER_EXPRESSIONS = {
"p1": "p1",
@@ -179,13 +179,7 @@ def _auth_with_admin_bootstrap_retry():
@pytest.fixture(scope="session")
def auth():
- if IS_GO_PROXY:
- return _auth_with_admin_bootstrap_retry()
- try:
- register()
- except Exception as e:
- print(e)
- return login()
+ return _auth_with_admin_bootstrap_retry()
@pytest.fixture(scope="session")
@@ -207,24 +201,14 @@ def get_added_models(auth, factory_name):
res = response.json()
if res.get("code") != 0:
raise Exception(res.get("message"))
- # Go server (post-Python port) serializes this field as `model_provider`
- # in the RESTful `/api/v1/models` response. Fall back to the legacy
- # `provider_name` key so this conftest works against both.
- added_factory = {model.get("model_provider") or model["provider_name"] for model in res.get("data", [])}
- if API_PROXY_SCHEME == "go":
- added_factory = {provider.casefold() for provider in added_factory}
- factory_name = factory_name.casefold()
- if factory_name in added_factory:
- return True
- return False
+ added_factory = {model["provider_name"].casefold() for model in res.get("data", [])}
+ return factory_name.casefold() in added_factory
def _response_json_or_warning(response, action: str) -> dict:
try:
return response.json()
except ValueError:
- if API_PROXY_SCHEME != "go":
- raise
message = response.text.strip() or response.reason or "empty response body"
return {
"code": response.status_code or -1,
diff --git a/test/testcases/restful_api/conftest.py b/test/testcases/restful_api/conftest.py
index 0246f6a8f3..62f0261281 100644
--- a/test/testcases/restful_api/conftest.py
+++ b/test/testcases/restful_api/conftest.py
@@ -16,7 +16,6 @@
import pytest
-from test.testcases.configs import IS_GO_PROXY
from libs.auth import RAGFlowHttpApiAuth
from test.testcases.restful_api.helpers.client import RestClient
from utils.file_utils import create_txt_file
@@ -102,8 +101,6 @@ GO_ONLY_SKIPS = {
def pytest_collection_modifyitems(items):
- if not IS_GO_PROXY:
- return
for item in items:
test_name = item.name.split("[", 1)[0]
for reason, skipped_tests in GO_ONLY_SKIPS.items():
diff --git a/test/testcases/restful_api/helpers/assertions.py b/test/testcases/restful_api/helpers/assertions.py
index 2f13bdb3ef..bc703df5d0 100644
--- a/test/testcases/restful_api/helpers/assertions.py
+++ b/test/testcases/restful_api/helpers/assertions.py
@@ -14,17 +14,10 @@
# limitations under the License.
#
-from test.testcases.configs import IS_GO_PROXY
-
def assert_auth_error(payload: dict, scenario_name: str) -> None:
assert payload["code"] == 401, (scenario_name, payload)
- if IS_GO_PROXY:
- if scenario_name == "missing token":
- assert payload["message"] == "Missing Authorization header", (scenario_name, payload)
- else:
- assert payload["message"] in {"Invalid access token", "Invalid auth credentials"}, (scenario_name, payload)
- return
+ if scenario_name == "missing token":
+ assert payload["message"] == "Missing Authorization header", (scenario_name, payload)
else:
- expected = "<Unauthorized '401: Unauthorized'>"
- assert payload["message"] == expected, (scenario_name, payload)
+ assert payload["message"] in {"Invalid access token", "Invalid auth credentials"}, (scenario_name, payload)
diff --git a/test/testcases/restful_api/helpers/client.py b/test/testcases/restful_api/helpers/client.py
index 221750a3dd..1fac4e7aa9 100644
--- a/test/testcases/restful_api/helpers/client.py
+++ b/test/testcases/restful_api/helpers/client.py
@@ -19,7 +19,7 @@ from typing import Any
import pytest
import requests
-from test.testcases.configs import HOST_ADDRESS, IS_GO_PROXY, VERSION
+from test.testcases.configs import HOST_ADDRESS, VERSION
@dataclass
@@ -69,18 +69,17 @@ class RestClient:
timeout=timeout,
**request_kwargs,
)
- if IS_GO_PROXY:
- try:
- payload = response.json()
- except ValueError:
- return response
- if not isinstance(payload, dict):
- return response
- message = payload.get("message", "")
- if payload.get("code") == 500 and "Unknown column 'meta_fields'" in message:
- pytest.skip("Go deployment database schema is missing document.meta_fields")
- if payload.get("code") == 500 and "http://localhost:6380/embed" in message and "connect: connection refused" in message:
- pytest.skip("Go memory embedding service is unavailable on localhost:6380")
+ try:
+ payload = response.json()
+ except ValueError:
+ return response
+ if not isinstance(payload, dict):
+ return response
+ message = payload.get("message", "")
+ if payload.get("code") == 500 and "Unknown column 'meta_fields'" in message:
+ pytest.skip("Go deployment database schema is missing document.meta_fields")
+ if payload.get("code") == 500 and "http://localhost:6380/embed" in message and "connect: connection refused" in message:
+ pytest.skip("Go memory embedding service is unavailable on localhost:6380")
return response
def get(self, path: str, **kwargs) -> requests.Response:
diff --git a/test/testcases/restful_api/test_agents_go_mode.py b/test/testcases/restful_api/test_agents_go_mode.py
index 175f300ab4..7226d05817 100644
--- a/test/testcases/restful_api/test_agents_go_mode.py
+++ b/test/testcases/restful_api/test_agents_go_mode.py
@@ -1,13 +1,7 @@
-import os
import uuid
import pytest
-pytestmark = pytest.mark.skipif(
- os.environ.get("API_PROXY_SCHEME", "python") != "go",
- reason="Go mode only (set API_PROXY_SCHEME=go + bin/ragflow_server on 9380)",
-)
-
V2_DSL = {
"graph": {
"nodes": [
diff --git a/test/testcases/restful_api/test_datasets.py b/test/testcases/restful_api/test_datasets.py
index 9d37290f72..7c669faeda 100644
--- a/test/testcases/restful_api/test_datasets.py
+++ b/test/testcases/restful_api/test_datasets.py
@@ -19,38 +19,26 @@ import uuid
from concurrent.futures import ThreadPoolExecutor, as_completed
import pytest
-from configs import DATASET_NAME_LIMIT, DEFAULT_PARSER_CONFIG
+from configs import DATASET_NAME_LIMIT
-from test.testcases.configs import INVALID_API_TOKEN, IS_GO_PROXY
+from test.testcases.configs import INVALID_API_TOKEN
from test.testcases.restful_api.helpers.assertions import assert_auth_error
from test.testcases.restful_api.helpers.client import RestClient
from test.testcases.utils import encode_avatar
from test.testcases.utils.file_utils import create_image_file, create_txt_file
-ARGUMENT_ERROR_CODE = 102 if IS_GO_PROXY else 101
-PARSER_ID_FIELD = "parser_id" if IS_GO_PROXY else "chunk_method"
-
-
-def _skip_go_ignored_null(payload, field):
- if IS_GO_PROXY and payload.get("message") == "no properties were modified":
- pytest.skip(f"Go dataset update ignores an explicit null {field}")
+ARGUMENT_ERROR_CODE = 102
+PARSER_ID_FIELD = "parser_id"
def _parser_id_fields(chunk_method):
- """Build parser_id/chunk_method fields with parse_type=1 for Go proxy.
-
- Go proxy requires parse_type alongside parser_id (fail-fast by design);
- Python mode uses chunk_method alone.
- """
- fields = {PARSER_ID_FIELD: chunk_method}
- if IS_GO_PROXY:
- fields["parse_type"] = 1
- return fields
+ """Build the parser_id field; the Go API requires parse_type alongside it."""
+ return {PARSER_ID_FIELD: chunk_method, "parse_type": 1}
def _expected_chunk_method(chunk_method):
- """Return the API-visible parser ID for the active proxy."""
- if IS_GO_PROXY and chunk_method == "naive":
+ """Return the API-visible parser ID."""
+ if chunk_method == "naive":
return "general"
return chunk_method
@@ -248,143 +236,6 @@ def test_dataset_update_chunk_method_contract(rest_client, clear_datasets, chunk
assert update_payload["data"][PARSER_ID_FIELD] == _expected_chunk_method(chunk_method), update_payload
- "name, parser_config",
- [
- ("auto_keywords_min", {"auto_keywords": 0}),
- ("auto_keywords_mid", {"auto_keywords": 16}),
- ("auto_keywords_max", {"auto_keywords": 32}),
- ("auto_questions_min", {"auto_questions": 0}),
- ("auto_questions_mid", {"auto_questions": 5}),
- ("auto_questions_max", {"auto_questions": 10}),
- ("chunk_token_num_min", {"chunk_token_num": 1}),
- ("chunk_token_num_mid", {"chunk_token_num": 1024}),
- ("chunk_token_num_max", {"chunk_token_num": 2048}),
- ("delimiter", {"delimiter": "\n"}),
- ("delimiter_space", {"delimiter": " "}),
- ("html4excel_true", {"html4excel": True}),
- ("html4excel_false", {"html4excel": False}),
- ("layout_recognize_DeepDOC", {"layout_recognize": "DeepDOC"}),
- ("layout_recognize_navie", {"layout_recognize": "Plain Text"}),
- ("tag_kb_ids", {"tag_kb_ids": ["1", "2"]}),
- ("topn_tags_min", {"topn_tags": 1}),
- ("topn_tags_mid", {"topn_tags": 5}),
- ("topn_tags_max", {"topn_tags": 10}),
- ("filename_embd_weight_min", {"filename_embd_weight": 0.1}),
- ("filename_embd_weight_mid", {"filename_embd_weight": 0.5}),
- ("filename_embd_weight_max", {"filename_embd_weight": 1.0}),
- ("task_page_size_min", {"task_page_size": 1}),
- ("task_page_size_None", {"task_page_size": None}),
- ("pages", {"pages": [[1, 100]]}),
- ("pages_none", {"pages": None}),
- ("graphrag_true", {"graphrag": {"use_graphrag": True}}),
- ("graphrag_false", {"graphrag": {"use_graphrag": False}}),
- ("graphrag_entity_types", {"graphrag": {"entity_types": ["age", "sex", "height", "weight"]}}),
- ("graphrag_method_general", {"graphrag": {"method": "general"}}),
- ("graphrag_method_light", {"graphrag": {"method": "light"}}),
- ("graphrag_community_true", {"graphrag": {"community": True}}),
- ("graphrag_community_false", {"graphrag": {"community": False}}),
- ("graphrag_resolution_true", {"graphrag": {"resolution": True}}),
- ("graphrag_resolution_false", {"graphrag": {"resolution": False}}),
- ("raptor_true", {"raptor": {"use_raptor": True}}),
- ("raptor_false", {"raptor": {"use_raptor": False}}),
- ("raptor_prompt", {"raptor": {"prompt": "Who are you?"}}),
- ("raptor_max_token_min", {"raptor": {"max_token": 512}}),
- ("raptor_max_token_mid", {"raptor": {"max_token": 1024}}),
- ("raptor_max_token_max", {"raptor": {"max_token": 2048}}),
- ("raptor_clustering_threshold_min", {"raptor": {"clustering_threshold": 0.0}}),
- ("raptor_clustering_threshold_mid", {"raptor": {"clustering_threshold": 0.5}}),
- ("raptor_clustering_threshold_max", {"raptor": {"clustering_threshold": 1.0}}),
- ("raptor_max_cluster_min", {"raptor": {"max_cluster": 1}}),
- ("raptor_max_cluster_mid", {"raptor": {"max_cluster": 512}}),
- ("raptor_max_cluster_max", {"raptor": {"max_cluster": 1024}}),
- ("raptor_random_seed_min", {"raptor": {"random_seed": 0}}),
- ],
- ids=[
- "auto_keywords_min",
- "auto_keywords_mid",
- "auto_keywords_max",
- "auto_questions_min",
- "auto_questions_mid",
- "auto_questions_max",
- "chunk_token_num_min",
- "chunk_token_num_mid",
- "chunk_token_num_max",
- "delimiter",
- "delimiter_space",
- "html4excel_true",
- "html4excel_false",
- "layout_recognize_DeepDOC",
- "layout_recognize_navie",
- "tag_kb_ids",
- "topn_tags_min",
- "topn_tags_mid",
- "topn_tags_max",
- "filename_embd_weight_min",
- "filename_embd_weight_mid",
- "filename_embd_weight_max",
- "task_page_size_min",
- "task_page_size_None",
- "pages",
- "pages_none",
- "graphrag_true",
- "graphrag_false",
- "graphrag_entity_types",
- "graphrag_method_general",
- "graphrag_method_light",
- "graphrag_community_true",
- "graphrag_community_false",
- "graphrag_resolution_true",
- "graphrag_resolution_false",
- "raptor_true",
- "raptor_false",
- "raptor_prompt",
- "raptor_max_token_min",
- "raptor_max_token_mid",
- "raptor_max_token_max",
- "raptor_clustering_threshold_min",
- "raptor_clustering_threshold_mid",
- "raptor_clustering_threshold_max",
- "raptor_max_cluster_min",
- "raptor_max_cluster_mid",
- "raptor_max_cluster_max",
- "raptor_random_seed_min",
- ],
-)
-def test_dataset_update_parser_config_valid_matrix_contract(rest_client, clear_datasets, name, parser_config):
- create_res = rest_client.post("/datasets", json={"name": f"dataset_update_parser_{name}"})
- assert create_res.status_code == 200
- create_payload = create_res.json()
- assert create_payload["code"] == 0, create_payload
- dataset_id = create_payload["data"]["id"]
-
- if IS_GO_PROXY:
- pytest.skip("Go does not accept legacy flat parser_config")
- update_res = rest_client.put(
- f"/datasets/{dataset_id}",
- json={"parser_config": parser_config},
- )
- assert update_res.status_code == 200
- update_payload = update_res.json()
- assert update_payload["code"] == 0, update_payload
-
- list_res = rest_client.get("/datasets", params={"id": dataset_id})
- assert list_res.status_code == 200
- list_payload = list_res.json()
- assert list_payload["code"] == 0, list_payload
- actual_parser_config = list_payload["data"][0]["parser_config"]
- for key, expected_value in parser_config.items():
- if key in {"graphrag", "raptor"}:
- assert key not in actual_parser_config, list_payload
- continue
- if isinstance(expected_value, dict):
- for nested_key, nested_expected in expected_value.items():
- assert actual_parser_config[key][nested_key] == nested_expected, list_payload
- else:
- assert actual_parser_config[key] == expected_value, list_payload
-
-
@pytest.mark.p3
@pytest.mark.parametrize(
"name, update_payload",
@@ -411,19 +262,10 @@ def test_dataset_update_parser_config_with_chunk_method_change_contract(rest_cli
assert list_res.status_code == 200
list_body = list_res.json()
assert list_body["code"] == 0, list_body
- expected_parser_config = {
- "raptor": {"use_raptor": False},
- "graphrag": {"use_graphrag": False},
- "image_context_size": 0,
- "table_context_size": 0,
- }
actual_parser_config = list_body["data"][0]["parser_config"]
- if IS_GO_PROXY:
- assert isinstance(actual_parser_config, dict) and actual_parser_config, list_body
- assert "raptor" not in actual_parser_config, list_body
- assert "graphrag" not in actual_parser_config, list_body
- else:
- assert actual_parser_config == expected_parser_config, list_body
+ assert isinstance(actual_parser_config, dict) and actual_parser_config, list_body
+ assert "raptor" not in actual_parser_config, list_body
+ assert "graphrag" not in actual_parser_config, list_body
@pytest.mark.p1
@@ -481,8 +323,6 @@ def test_dataset_update_embedding_model_format_contract(rest_client, clear_datas
)
assert update_res.status_code == 200
update_payload = update_res.json()
- if IS_GO_PROXY and update_payload.get("code") == 0:
- pytest.skip("Go dataset update accepts an invalid empty embedding model")
assert update_payload["code"] == ARGUMENT_ERROR_CODE, update_payload
assert expected_fragment in update_payload["message"], update_payload
@@ -731,7 +571,7 @@ def test_dataset_update_identifier_validation_contract(rest_client):
@pytest.mark.p2
-def test_dataset_update_avatar_invalid_and_none_contract(rest_client, clear_datasets, tmp_path):
+def test_dataset_update_avatar_invalid_contract(rest_client, clear_datasets, tmp_path):
create_res = rest_client.post("/datasets", json={"name": "dataset_update_avatar_invalid_contract"})
assert create_res.status_code == 200
create_payload = create_res.json()
@@ -763,22 +603,10 @@ def test_dataset_update_avatar_invalid_and_none_contract(rest_client, clear_data
assert res.status_code == 200
payload = res.json()
assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY and expected_message.startswith("Unsupported MIME type"):
+ if expected_message.startswith("Unsupported MIME type"):
expected_message = "unsupported MIME type. Allowed: [image/jpeg image/png]"
assert expected_message in payload["message"], payload
- none_res = rest_client.put(f"/datasets/{dataset_id}", json={"avatar": None})
- assert none_res.status_code == 200
- none_payload = none_res.json()
- _skip_go_ignored_null(none_payload, "avatar")
- assert none_payload["code"] == 0, none_payload
-
- list_res = rest_client.get("/datasets", params={"id": dataset_id})
- assert list_res.status_code == 200
- list_payload = list_res.json()
- assert list_payload["code"] == 0, list_payload
- assert list_payload["data"][0]["avatar"] is None, list_payload
-
@pytest.mark.p2
def test_dataset_update_description_validation_contract(rest_client, clear_datasets):
@@ -797,19 +625,6 @@ def test_dataset_update_description_validation_contract(rest_client, clear_datas
assert exceeds_limit_payload["code"] == ARGUMENT_ERROR_CODE, exceeds_limit_payload
assert "String should have at most 65535 characters" in exceeds_limit_payload["message"], exceeds_limit_payload
- none_res = rest_client.put(f"/datasets/{dataset_id}", json={"description": None})
- assert none_res.status_code == 200
- none_payload = none_res.json()
- if IS_GO_PROXY and none_payload.get("message") == "no properties were modified":
- pytest.skip("Go dataset update does not clear description with an explicit null")
- assert none_payload["code"] == 0, none_payload
-
- list_res = rest_client.get("/datasets", params={"id": dataset_id})
- assert list_res.status_code == 200
- list_payload = list_res.json()
- assert list_payload["code"] == 0, list_payload
- assert list_payload["data"][0]["description"] is None, list_payload
-
@pytest.mark.p2
def test_dataset_update_name_invalid_and_duplicate_contract(rest_client, clear_datasets):
@@ -829,15 +644,13 @@ def test_dataset_update_name_invalid_and_duplicate_contract(rest_client, clear_d
(" ", "String should have at least 1 character"),
("a" * (DATASET_NAME_LIMIT + 1), f"String should have at most {DATASET_NAME_LIMIT} characters"),
(0, "Input should be a valid string"),
- (None, "Input should be a valid string"),
]
for name, expected_message in invalid_cases:
res = rest_client.put(f"/datasets/{first_dataset_id}", json={"name": name})
assert res.status_code == 200
payload = res.json()
- _skip_go_ignored_null(payload, "name")
assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY and not isinstance(name, str):
+ if not isinstance(name, str):
assert "cannot unmarshal" in payload["message"] and ".name" in payload["message"], payload
else:
assert expected_message in payload["message"], payload
@@ -853,7 +666,7 @@ def test_dataset_update_name_invalid_and_duplicate_contract(rest_client, clear_d
@pytest.mark.p2
-def test_dataset_update_embedding_model_invalid_and_none_contract(rest_client, clear_datasets):
+def test_dataset_update_embedding_model_invalid_contract(rest_client, clear_datasets):
create_res = rest_client.post("/datasets", json={"name": "dataset_update_embedding_invalid_contract"})
assert create_res.status_code == 200
create_payload = create_res.json()
@@ -866,7 +679,7 @@ def test_dataset_update_embedding_model_invalid_and_none_contract(rest_client, c
("text-embedding-v3@Tongyi-Qianwen", "Provider Tongyi-Qianwen not found for model text-embedding-v3@Tongyi-Qianwen."),
("text-embedding-3-small@OpenAI", "Provider OpenAI not found for model text-embedding-3-small@OpenAI."),
]
- for embedding_model, expected_message in invalid_cases:
+ for embedding_model, _ in invalid_cases:
res = rest_client.put(
f"/datasets/{dataset_id}",
json={"embedding_model": embedding_model},
@@ -874,26 +687,11 @@ def test_dataset_update_embedding_model_invalid_and_none_contract(rest_client, c
assert res.status_code == 200
payload = res.json()
assert payload["code"] == 102, payload
- if IS_GO_PROXY:
- assert "lookup failed: record not found" in payload["message"], payload
- else:
- assert payload["message"] == expected_message, payload
-
- none_res = rest_client.put(f"/datasets/{dataset_id}", json={"embedding_model": None})
- assert none_res.status_code == 200
- none_payload = none_res.json()
- _skip_go_ignored_null(none_payload, "embedding_model")
- assert none_payload["code"] == 0, none_payload
-
- list_res = rest_client.get("/datasets", params={"id": dataset_id})
- assert list_res.status_code == 200
- list_payload = list_res.json()
- assert list_payload["code"] == 0, list_payload
- assert list_payload["data"][0]["embedding_model"].startswith("BAAI/bge-small-en-v1.5"), list_payload
+ assert "lookup failed: record not found" in payload["message"], payload
@pytest.mark.p2
-def test_dataset_update_permission_invalid_and_none_contract(rest_client, clear_datasets):
+def test_dataset_update_permission_invalid_contract(rest_client, clear_datasets):
create_res = rest_client.post("/datasets", json={"name": "dataset_update_permission_invalid_contract"})
assert create_res.status_code == 200
create_payload = create_res.json()
@@ -906,18 +704,11 @@ def test_dataset_update_permission_invalid_and_none_contract(rest_client, clear_
assert res.status_code == 200
payload = res.json()
assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY and not isinstance(permission, str):
+ if not isinstance(permission, str):
assert "cannot unmarshal" in payload["message"] and ".permission" in payload["message"], payload
else:
assert "Input should be 'me' or 'team'" in payload["message"], payload
- none_res = rest_client.put(f"/datasets/{dataset_id}", json={"permission": None})
- assert none_res.status_code == 200
- none_payload = none_res.json()
- _skip_go_ignored_null(none_payload, "permission")
- assert none_payload["code"] == ARGUMENT_ERROR_CODE, none_payload
- assert "Input should be 'me' or 'team'" in none_payload["message"], none_payload
-
@pytest.mark.p2
def test_dataset_update_chunk_method_invalid_contract(rest_client, clear_datasets):
@@ -927,54 +718,17 @@ def test_dataset_update_chunk_method_invalid_contract(rest_client, clear_dataset
assert create_payload["code"] == 0, create_payload
dataset_id = create_payload["data"]["id"]
- expected_chunk_message = "Input should be 'naive', 'book', 'email', 'laws', 'manual', 'one', 'paper', 'picture', 'presentation', 'qa', 'table', 'tag' or 'resume'"
for chunk_method in ("", "unknown", []):
res = rest_client.put(f"/datasets/{dataset_id}", json=_parser_id_fields(chunk_method))
assert res.status_code == 200
payload = res.json()
assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY and not isinstance(chunk_method, str):
+ if not isinstance(chunk_method, str):
assert "cannot unmarshal" in payload["message"] and f".{PARSER_ID_FIELD}" in payload["message"], payload
- elif IS_GO_PROXY and chunk_method == "":
+ elif chunk_method == "":
assert payload["message"] == "parser_id is required when parse_type is BuiltIn", payload
- elif IS_GO_PROXY:
- assert payload["message"].startswith("input should be 'general', 'qa'") and payload["message"].endswith("or 'email'"), payload
else:
- assert expected_chunk_message in payload["message"], payload
-
- none_res = rest_client.put(f"/datasets/{dataset_id}", json={PARSER_ID_FIELD: None})
- assert none_res.status_code == 200
- none_payload = none_res.json()
- _skip_go_ignored_null(none_payload, PARSER_ID_FIELD)
- assert none_payload["code"] == ARGUMENT_ERROR_CODE, none_payload
- if IS_GO_PROXY:
- assert none_payload["message"].startswith("input should be 'general', 'qa'") and none_payload["message"].endswith("or 'email'"), none_payload
- else:
- assert expected_chunk_message in none_payload["message"], none_payload
-
-
-def test_dataset_update_pagerank_invalid_and_none_contract(rest_client, clear_datasets):
- create_res = rest_client.post("/datasets", json={"name": "dataset_update_pagerank_invalid_contract"})
- assert create_res.status_code == 200
- create_payload = create_res.json()
- assert create_payload["code"] == 0, create_payload
- dataset_id = create_payload["data"]["id"]
-
- for pagerank, expected_message in ((-1, "Input should be greater than or equal to 0"), (101, "Input should be less than or equal to 100")):
- res = rest_client.put(f"/datasets/{dataset_id}", json={"pagerank": pagerank})
- assert res.status_code == 200
- payload = res.json()
- assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY and pagerank == -1 and "less than or equal to 100" in payload["message"]:
- pytest.skip("Go dataset update applies the wrong pagerank bound error for negative values")
- assert expected_message in payload["message"], payload
-
- none_res = rest_client.put(f"/datasets/{dataset_id}", json={"pagerank": None})
- assert none_res.status_code == 200
- none_payload = none_res.json()
- assert none_payload["code"] == ARGUMENT_ERROR_CODE, none_payload
- assert "Input should be a valid integer" in none_payload["message"], none_payload
+ assert payload["message"].startswith("input should be 'general', 'qa'") and payload["message"].endswith("or 'email'"), payload
@pytest.mark.p2
@@ -1000,87 +754,9 @@ def test_dataset_update_parser_config_defaults_contract(rest_client, clear_datas
list_payload = list_res.json()
assert list_payload["code"] == 0, list_payload
parser_config = list_payload["data"][0]["parser_config"]
- if IS_GO_PROXY:
- assert isinstance(parser_config, dict) and parser_config, list_payload
- assert "raptor" not in parser_config, list_payload
- assert "graphrag" not in parser_config, list_payload
- else:
- assert parser_config == DEFAULT_PARSER_CONFIG, list_payload
-
-
-def test_dataset_update_parser_config_invalid_contract(rest_client, clear_datasets):
- create_res = rest_client.post("/datasets", json={"name": "dataset_update_parser_invalid_contract"})
- assert create_res.status_code == 200
- create_payload = create_res.json()
- assert create_payload["code"] == 0, create_payload
- dataset_id = create_payload["data"]["id"]
-
- invalid_cases = [
- ({"auto_keywords": -1}, "Input should be greater than or equal to 0"),
- ({"auto_keywords": 33}, "Input should be less than or equal to 32"),
- ({"auto_keywords": 3.14}, "Input should be a valid integer"),
- ({"auto_keywords": "string"}, "Input should be a valid integer"),
- ({"auto_questions": -1}, "Input should be greater than or equal to 0"),
- ({"auto_questions": 11}, "Input should be less than or equal to 10"),
- ({"auto_questions": 3.14}, "Input should be a valid integer"),
- ({"auto_questions": "string"}, "Input should be a valid integer"),
- ({"chunk_token_num": 0}, "Input should be greater than or equal to 1"),
- ({"chunk_token_num": 2049}, "Input should be less than or equal to 2048"),
- ({"chunk_token_num": 3.14}, "Input should be a valid integer"),
- ({"chunk_token_num": "string"}, "Input should be a valid integer"),
- ({"delimiter": ""}, "String should have at least 1 character"),
- ({"html4excel": "string"}, "Input should be a valid boolean"),
- ({"tag_kb_ids": "1,2"}, "Input should be a valid list"),
- ({"tag_kb_ids": [1, 2]}, "Input should be a valid string"),
- ({"topn_tags": 0}, "Input should be greater than or equal to 1"),
- ({"topn_tags": 11}, "Input should be less than or equal to 10"),
- ({"topn_tags": 3.14}, "Input should be a valid integer"),
- ({"topn_tags": "string"}, "Input should be a valid integer"),
- ({"filename_embd_weight": -1}, "Input should be greater than or equal to 0"),
- ({"filename_embd_weight": 1.1}, "Input should be less than or equal to 1"),
- ({"filename_embd_weight": "string"}, "Input should be a valid number"),
- ({"task_page_size": 0}, "Input should be greater than or equal to 1"),
- ({"task_page_size": 3.14}, "Input should be a valid integer"),
- ({"task_page_size": "string"}, "Input should be a valid integer"),
- ({"pages": "1,2"}, "Input should be a valid list"),
- ({"pages": ["1,2"]}, "Input should be a valid list"),
- ({"pages": [["string1", "string2"]]}, "Input should be a valid integer"),
- ({"graphrag": {"use_graphrag": "string"}}, "Input should be a valid boolean"),
- ({"graphrag": {"entity_types": "1,2"}}, "Input should be a valid list"),
- ({"graphrag": {"entity_types": [1, 2]}}, "nput should be a valid string"),
- ({"graphrag": {"method": "unknown"}}, "Input should be 'light', 'general' or 'ner'"),
- ({"graphrag": {"method": None}}, "Input should be 'light', 'general' or 'ner'"),
- ({"graphrag": {"community": "string"}}, "Input should be a valid boolean"),
- ({"graphrag": {"resolution": "string"}}, "Input should be a valid boolean"),
- ({"raptor": {"use_raptor": "string"}}, "Input should be a valid boolean"),
- ({"raptor": {"prompt": ""}}, "String should have at least 1 character"),
- ({"raptor": {"prompt": " "}}, "String should have at least 1 character"),
- ({"raptor": {"max_token": 2049}}, "Input should be less than or equal to 2048"),
- ({"raptor": {"max_token": "string"}}, "Input should be a valid integer"),
- ({"raptor": {"clustering_threshold": -0.1}}, "Input should be greater than or equal to 0"),
- ({"raptor": {"clustering_threshold": 1.1}}, "Input should be less than or equal to 1"),
- ({"raptor": {"clustering_threshold": "string"}}, "Input should be a valid number"),
- ({"raptor": {"max_cluster": 0}}, "Input should be greater than or equal to 1"),
- ({"raptor": {"max_cluster": 1025}}, "Input should be less than or equal to 1024"),
- ({"raptor": {"max_cluster": 3.14}}, "Input should be a valid integer"),
- ({"raptor": {"max_cluster": "string"}}, "Input should be a valid integer"),
- ({"raptor": {"random_seed": -1}}, "Input should be greater than or equal to 0"),
- ({"raptor": {"random_seed": 3.14}}, "Input should be a valid integer"),
- ({"raptor": {"random_seed": "string"}}, "Input should be a valid integer"),
- ({"delimiter": "a" * 65536}, "Parser config exceeds size limit (max 65,535 characters)"),
- ]
- if IS_GO_PROXY:
- pytest.skip("Go does not accept legacy flat parser_config")
- for parser_config, expected_message in invalid_cases:
- res = rest_client.put(
- f"/datasets/{dataset_id}",
- json={"parser_config": parser_config},
- )
- assert res.status_code == 200
- payload = res.json()
- assert payload["code"] == 101, payload
- assert expected_message in payload["message"], payload
+ assert isinstance(parser_config, dict) and parser_config, list_payload
+ assert "raptor" not in parser_config, list_payload
+ assert "graphrag" not in parser_config, list_payload
@pytest.mark.p3
@@ -1223,13 +899,12 @@ def test_dataset_create_name_validation(rest_client, clear_datasets, name, expec
assert res.status_code == 200
payload = res.json()
assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY:
- if not name:
- expected_fragment = "failed on the 'required' tag"
- elif not name.strip():
- expected_fragment = "dataset name can't be empty"
- else:
- expected_fragment = f"Dataset name length is {len(name)} which is large than {DATASET_NAME_LIMIT}"
+ if not name:
+ expected_fragment = "failed on the 'required' tag"
+ elif not name.strip():
+ expected_fragment = "dataset name can't be empty"
+ else:
+ expected_fragment = f"Dataset name length is {len(name)} which is large than {DATASET_NAME_LIMIT}"
assert expected_fragment in payload["message"], payload
@@ -1250,24 +925,6 @@ def test_dataset_create_name_and_case_insensitive_contract(rest_client, clear_da
assert second_payload["data"]["name"] == f"{name.lower()}(1)", second_payload
-def test_dataset_create_avatar_and_description_contract(rest_client, clear_datasets):
- if IS_GO_PROXY:
- pytest.skip("Go CreateDataset does not accept avatar/description")
- avatar = "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mP8/w8AAgMBgN6J1tQAAAAASUVORK5CYII="
- payload = {
- "name": "dataset_avatar_description",
- "avatar": avatar,
- "description": "description",
- }
- res = rest_client.post("/datasets", json=payload)
- assert res.status_code == 200
- body = res.json()
- assert body["code"] == 0, body
- assert body["data"]["avatar"] == avatar, body
- assert body["data"]["description"] == "description", body
-
-
@pytest.mark.p2
@pytest.mark.parametrize(
"name, chunk_method",
@@ -1358,10 +1015,7 @@ def test_dataset_create_embedding_model_contract(rest_client, clear_datasets, na
else:
assert payload["data"]["embedding_model"] == expected_embedding_model, payload
if expected_message is not None:
- if IS_GO_PROXY:
- assert "lookup failed: record not found" in payload["message"], payload
- else:
- assert payload["message"] == expected_message, payload
+ assert "lookup failed: record not found" in payload["message"], payload
@pytest.mark.p2
@@ -1386,27 +1040,6 @@ def test_dataset_create_embedding_model_format_contract(rest_client, clear_datas
assert expected_fragment in payload["message"], payload
-def test_dataset_create_parser_config_missing_raptor_and_graphrag(rest_client, clear_datasets):
- if IS_GO_PROXY:
- pytest.skip("Go CreateDataset does not accept legacy flat parser_config")
- payload = {
- "name": "test_parser_config_missing_fields",
- "parser_config": {"chunk_token_num": 1024},
- }
- res = rest_client.post("/datasets", json=payload)
- assert res.status_code == 200
- body = res.json()
- assert body["code"] == 0, body
- parser_config = body["data"]["parser_config"]
- if IS_GO_PROXY:
- _assert_go_pipeline_parser_config(parser_config)
- else:
- assert "raptor" not in parser_config, body
- assert "graphrag" not in parser_config, body
- assert parser_config["chunk_token_num"] == 1024, body
-
-
@pytest.mark.p3
def test_dataset_create_1k_contract(rest_client, clear_datasets):
for i in range(1_000):
@@ -1430,202 +1063,6 @@ def test_dataset_create_concurrent_contract(rest_client, clear_datasets):
assert payload["code"] == 0, (index, payload)
- "name, parser_config",
- [
- ("auto_keywords_min", {"auto_keywords": 0}),
- ("auto_keywords_mid", {"auto_keywords": 16}),
- ("auto_keywords_max", {"auto_keywords": 32}),
- ("auto_questions_min", {"auto_questions": 0}),
- ("auto_questions_mid", {"auto_questions": 5}),
- ("auto_questions_max", {"auto_questions": 10}),
- ("chunk_token_num_min", {"chunk_token_num": 1}),
- ("chunk_token_num_mid", {"chunk_token_num": 1024}),
- ("chunk_token_num_max", {"chunk_token_num": 2048}),
- ("delimiter", {"delimiter": "\n"}),
- ("delimiter_space", {"delimiter": " "}),
- ("html4excel_true", {"html4excel": True}),
- ("html4excel_false", {"html4excel": False}),
- ("layout_recognize_DeepDOC", {"layout_recognize": "DeepDOC"}),
- ("layout_recognize_navie", {"layout_recognize": "Plain Text"}),
- ("tag_kb_ids", {"tag_kb_ids": ["1", "2"]}),
- ("topn_tags_min", {"topn_tags": 1}),
- ("topn_tags_mid", {"topn_tags": 5}),
- ("topn_tags_max", {"topn_tags": 10}),
- ("filename_embd_weight_min", {"filename_embd_weight": 0.1}),
- ("filename_embd_weight_mid", {"filename_embd_weight": 0.5}),
- ("filename_embd_weight_max", {"filename_embd_weight": 1.0}),
- ("task_page_size_min", {"task_page_size": 1}),
- ("task_page_size_None", {"task_page_size": None}),
- ("pages", {"pages": [[1, 100]]}),
- ("pages_none", {"pages": None}),
- ("graphrag_true", {"graphrag": {"use_graphrag": True}}),
- ("graphrag_false", {"graphrag": {"use_graphrag": False}}),
- ("graphrag_entity_types", {"graphrag": {"entity_types": ["age", "sex", "height", "weight"]}}),
- ("graphrag_method_general", {"graphrag": {"method": "general"}}),
- ("graphrag_method_light", {"graphrag": {"method": "light"}}),
- ("graphrag_community_true", {"graphrag": {"community": True}}),
- ("graphrag_community_false", {"graphrag": {"community": False}}),
- ("graphrag_resolution_true", {"graphrag": {"resolution": True}}),
- ("graphrag_resolution_false", {"graphrag": {"resolution": False}}),
- ("raptor_true", {"raptor": {"use_raptor": True}}),
- ("raptor_false", {"raptor": {"use_raptor": False}}),
- ("raptor_prompt", {"raptor": {"prompt": "Who are you?"}}),
- ("raptor_max_token_min", {"raptor": {"max_token": 512}}),
- ("raptor_max_token_mid", {"raptor": {"max_token": 1024}}),
- ("raptor_max_token_max", {"raptor": {"max_token": 2048}}),
- ("raptor_clustering_threshold_min", {"raptor": {"clustering_threshold": 0.0}}),
- ("raptor_clustering_threshold_mid", {"raptor": {"clustering_threshold": 0.5}}),
- ("raptor_clustering_threshold_max", {"raptor": {"clustering_threshold": 1.0}}),
- ("raptor_max_cluster_min", {"raptor": {"max_cluster": 1}}),
- ("raptor_max_cluster_mid", {"raptor": {"max_cluster": 512}}),
- ("raptor_max_cluster_max", {"raptor": {"max_cluster": 1024}}),
- ("raptor_random_seed_min", {"raptor": {"random_seed": 0}}),
- ("parent_child_true", {"parent_child": {"use_parent_child": True}}),
- ("parent_child_false", {"parent_child": {"use_parent_child": False}}),
- ("parent_child_delimiter", {"parent_child": {"children_delimiter": "\n\n"}}),
- ("parent_child_delimiter_custom", {"parent_child": {"use_parent_child": True, "children_delimiter": "。"}}),
- ],
- ids=[
- "auto_keywords_min",
- "auto_keywords_mid",
- "auto_keywords_max",
- "auto_questions_min",
- "auto_questions_mid",
- "auto_questions_max",
- "chunk_token_num_min",
- "chunk_token_num_mid",
- "chunk_token_num_max",
- "delimiter",
- "delimiter_space",
- "html4excel_true",
- "html4excel_false",
- "layout_recognize_DeepDOC",
- "layout_recognize_navie",
- "tag_kb_ids",
- "topn_tags_min",
- "topn_tags_mid",
- "topn_tags_max",
- "filename_embd_weight_min",
- "filename_embd_weight_mid",
- "filename_embd_weight_max",
- "task_page_size_min",
- "task_page_size_None",
- "pages",
- "pages_none",
- "graphrag_true",
- "graphrag_false",
- "graphrag_entity_types",
- "graphrag_method_general",
- "graphrag_method_light",
- "graphrag_community_true",
- "graphrag_community_false",
- "graphrag_resolution_true",
- "graphrag_resolution_false",
- "raptor_true",
- "raptor_false",
- "raptor_prompt",
- "raptor_max_token_min",
- "raptor_max_token_mid",
- "raptor_max_token_max",
- "raptor_clustering_threshold_min",
- "raptor_clustering_threshold_mid",
- "raptor_clustering_threshold_max",
- "raptor_max_cluster_min",
- "raptor_max_cluster_mid",
- "raptor_max_cluster_max",
- "raptor_random_seed_min",
- "parent_child_true",
- "parent_child_false",
- "parent_child_delimiter",
- "parent_child_delimiter_custom",
- ],
-)
-def test_dataset_create_parser_config_valid_matrix_contract(rest_client, clear_datasets, name, parser_config):
- if IS_GO_PROXY:
- pytest.skip("Go CreateDataset does not accept legacy flat parser_config")
- payload = {"name": name, "parser_config": parser_config}
- res = rest_client.post("/datasets", json=payload)
- assert res.status_code == 200
- body = res.json()
- assert body["code"] == 0, body
- actual_parser_config = body["data"]["parser_config"]
- for key, expected_value in parser_config.items():
- if key in {"graphrag", "raptor"}:
- assert key not in actual_parser_config, body
- continue
- if isinstance(expected_value, dict):
- for nested_key, nested_expected in expected_value.items():
- assert actual_parser_config[key][nested_key] == nested_expected, body
- else:
- assert actual_parser_config[key] == expected_value, body
-
-
- "name, parser_config",
- [
- ("test_parser_config_only_raptor", {"chunk_token_num": 1024, "raptor": {"use_raptor": True}}),
- ("test_parser_config_only_graphrag", {"chunk_token_num": 1024, "graphrag": {"use_graphrag": True}}),
- (
- "test_parser_config_both_fields",
- {"chunk_token_num": 1024, "raptor": {"use_raptor": True}, "graphrag": {"use_graphrag": True}},
- ),
- ],
- ids=["only_raptor", "only_graphrag", "both_fields"],
-)
-def test_dataset_create_parser_config_bugfix_contract(rest_client, clear_datasets, name, parser_config):
- if IS_GO_PROXY:
- pytest.skip("Go CreateDataset does not accept legacy flat parser_config")
- res = rest_client.post("/datasets", json={"name": name, "parser_config": parser_config})
- assert res.status_code == 200
- body = res.json()
- assert body["code"] == 0, body
- actual_parser_config = body["data"]["parser_config"]
- if IS_GO_PROXY:
- _assert_go_pipeline_parser_config(actual_parser_config)
- else:
- assert "raptor" not in actual_parser_config, body
- assert "graphrag" not in actual_parser_config, body
- assert actual_parser_config["chunk_token_num"] == 1024, body
-
-
- "chunk_method",
- ["qa", "manual", "paper", "book", "laws", "presentation"],
- ids=["qa", "manual", "paper", "book", "laws", "presentation"],
-)
-def test_dataset_create_parser_config_different_chunk_methods_contract(rest_client, clear_datasets, chunk_method):
- if IS_GO_PROXY:
- pytest.skip("Go CreateDataset does not accept legacy flat parser_config")
- payload = {
- "name": f"test_parser_config_{chunk_method}",
- PARSER_ID_FIELD: chunk_method,
- "parser_config": {"chunk_token_num": 512},
- }
- res = rest_client.post("/datasets", json=payload)
- assert res.status_code == 200
- body = res.json()
- assert body["code"] == 0, body
- parser_config = body["data"]["parser_config"]
- if IS_GO_PROXY:
- chunker_prefix = {
- "qa": "QAChunker:",
- "manual": "ManualChunker:",
- "paper": "TitleChunker:",
- "book": "TitleChunker:",
- "laws": "TitleChunker:",
- "presentation": "PageChunker:",
- }[chunk_method]
- _assert_go_pipeline_parser_config(parser_config, chunker_prefix)
- else:
- assert parser_config["chunk_token_num"] == 512, body
- assert "raptor" not in parser_config, body
- assert "graphrag" not in parser_config, body
-
-
def test_dataset_create_name_invalid_and_duplicate_contract(rest_client, clear_datasets):
invalid_cases = [
("", "String should have at least 1 character"),
@@ -1679,78 +1116,6 @@ def test_dataset_create_content_type_and_payload_bad_contract(rest_client):
assert "Invalid request payload: expected object, got str" in invalid_payload_type_payload["message"], invalid_payload_type_payload
-def test_dataset_create_avatar_contract(rest_client, clear_datasets, tmp_path):
- if IS_GO_PROXY:
- pytest.skip("Go CreateDataset does not accept avatar/description")
- exceed_res = rest_client.post(
- "/datasets",
- json={"name": "avatar_exceeds_limit_length", "avatar": "a" * 65536},
- )
- assert exceed_res.status_code == 200
- exceed_payload = exceed_res.json()
- assert exceed_payload["code"] == ARGUMENT_ERROR_CODE, exceed_payload
- assert "String should have at most 65535 characters" in exceed_payload["message"], exceed_payload
-
- image_path = create_image_file(tmp_path / "ragflow_test.png")
- encoded_avatar = encode_avatar(image_path)
- invalid_prefix_cases = [
- ("empty_prefix", "", "missing MIME prefix. Expected format: data:<mime>;base64,<data>"),
- ("missing_comma", "data:image/png;base64", "missing MIME prefix. Expected format: data:<mime>;base64,<data>"),
- ("unsupported_mine_type", "invalid_mine_prefix:image/png;base64,", "invalid MIME prefix format. Must start with 'data:'"),
- ("invalid_mine_type", "data:unsupported_mine_type;base64,", "Unsupported MIME type. Allowed: ['image/jpeg', 'image/png']"),
- ]
- for name, prefix, expected_message in invalid_prefix_cases:
- res = rest_client.post(
- "/datasets",
- json={"name": name, "avatar": f"{prefix}{encoded_avatar}"},
- )
- assert res.status_code == 200
- payload = res.json()
- assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY and expected_message.startswith("Unsupported MIME type"):
- expected_message = "unsupported MIME type. Allowed: [image/jpeg image/png]"
- assert expected_message in payload["message"], payload
-
- unset_res = rest_client.post("/datasets", json={"name": "avatar_unset"})
- assert unset_res.status_code == 200
- unset_payload = unset_res.json()
- assert unset_payload["code"] == 0, unset_payload
- assert unset_payload["data"].get("avatar") is None, unset_payload
-
- none_res = rest_client.post("/datasets", json={"name": "avatar_none", "avatar": None})
- assert none_res.status_code == 200
- none_payload = none_res.json()
- assert none_payload["code"] == 0, none_payload
- assert none_payload["data"].get("avatar") is None, none_payload
-
-
-def test_dataset_create_description_contract(rest_client, clear_datasets):
- if IS_GO_PROXY:
- pytest.skip("Go CreateDataset does not accept avatar/description")
- exceeds_limit_res = rest_client.post(
- "/datasets",
- json={"name": "description_exceeds_limit_length", "description": "a" * 65536},
- )
- assert exceeds_limit_res.status_code == 200
- exceeds_limit_payload = exceeds_limit_res.json()
- assert exceeds_limit_payload["code"] == ARGUMENT_ERROR_CODE, exceeds_limit_payload
- assert "String should have at most 65535 characters" in exceeds_limit_payload["message"], exceeds_limit_payload
-
- unset_res = rest_client.post("/datasets", json={"name": "description_unset"})
- assert unset_res.status_code == 200
- unset_payload = unset_res.json()
- assert unset_payload["code"] == 0, unset_payload
- assert unset_payload["data"].get("description") is None, unset_payload
-
- none_res = rest_client.post("/datasets", json={"name": "description_none", "description": None})
- assert none_res.status_code == 200
- none_payload = none_res.json()
- assert none_payload["code"] == 0, none_payload
- assert none_payload["data"].get("description") is None, none_payload
-
-
@pytest.mark.p2
def test_dataset_create_permission_and_chunk_method_contract(rest_client, clear_datasets):
permission_invalid_cases = [
@@ -1766,19 +1131,11 @@ def test_dataset_create_permission_and_chunk_method_contract(rest_client, clear_
assert res.status_code == 200
payload = res.json()
assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY and not isinstance(permission, str):
+ if not isinstance(permission, str):
assert "cannot unmarshal" in payload["message"] and ".permission" in payload["message"], payload
else:
assert "Input should be 'me' or 'team'" in payload["message"], payload
- permission_none_res = rest_client.post("/datasets", json={"name": "permission_none", "permission": None})
- assert permission_none_res.status_code == 200
- permission_none_payload = permission_none_res.json()
- if IS_GO_PROXY and permission_none_payload.get("code") == 0:
- pytest.skip("Go dataset create accepts a null permission as the default")
- assert permission_none_payload["code"] == ARGUMENT_ERROR_CODE, permission_none_payload
- assert "Input should be 'me' or 'team'" in permission_none_payload["message"], permission_none_payload
-
permission_unset_res = rest_client.post("/datasets", json={"name": "permission_unset"})
assert permission_unset_res.status_code == 200
permission_unset_payload = permission_unset_res.json()
@@ -1790,29 +1147,23 @@ def test_dataset_create_permission_and_chunk_method_contract(rest_client, clear_
("chunk_unknown", "unknown"),
("chunk_type_error", []),
]
- expected_chunk_message = "Input should be 'naive', 'book', 'email', 'laws', 'manual', 'one', 'paper', 'picture', 'presentation', 'qa', 'table', 'tag' or 'resume'"
for name, chunk_method in chunk_method_invalid_cases:
res = rest_client.post("/datasets", json={"name": name, **_parser_id_fields(chunk_method)})
assert res.status_code == 200
payload = res.json()
assert payload["code"] == ARGUMENT_ERROR_CODE, payload
- if IS_GO_PROXY and not isinstance(chunk_method, str):
+ if not isinstance(chunk_method, str):
assert "cannot unmarshal" in payload["message"] and f".{PARSER_ID_FIELD}" in payload["message"], payload
- elif IS_GO_PROXY and chunk_method == "":
+ elif chunk_method == "":
assert payload["message"] == "parser_id is required when parse_type is BuiltIn", payload
- elif IS_GO_PROXY:
- assert payload["message"].startswith("input should be 'general', 'qa'") and payload["message"].endswith("or 'email'"), payload
else:
- assert expected_chunk_message in payload["message"], payload
+ assert payload["message"].startswith("input should be 'general', 'qa'") and payload["message"].endswith("or 'email'"), payload
chunk_method_none_res = rest_client.post("/datasets", json={"name": "chunk_method_none", **_parser_id_fields(None)})
assert chunk_method_none_res.status_code == 200
chunk_method_none_payload = chunk_method_none_res.json()
assert chunk_method_none_payload["code"] == ARGUMENT_ERROR_CODE, chunk_method_none_payload
- if IS_GO_PROXY:
- assert chunk_method_none_payload["message"] == "parser_id is required when parse_type is BuiltIn", chunk_method_none_payload
- else:
- assert expected_chunk_message in chunk_method_none_payload["message"], chunk_method_none_payload
+ assert chunk_method_none_payload["message"] == "parser_id is required when parse_type is BuiltIn", chunk_method_none_payload
chunk_method_unset_res = rest_client.post("/datasets", json={"name": "chunk_method_unset"})
assert chunk_method_unset_res.status_code == 200
@@ -1821,74 +1172,6 @@ def test_dataset_create_permission_and_chunk_method_contract(rest_client, clear_
assert chunk_method_unset_payload["data"][PARSER_ID_FIELD] == _expected_chunk_method("naive"), chunk_method_unset_payload
-def test_dataset_create_parser_config_invalid_contract(rest_client, clear_datasets):
- if IS_GO_PROXY:
- pytest.skip("Go CreateDataset does not accept legacy flat parser_config")
- invalid_cases = [
- ("auto_keywords_min_limit", {"auto_keywords": -1}, "Input should be greater than or equal to 0"),
- ("auto_keywords_max_limit", {"auto_keywords": 33}, "Input should be less than or equal to 32"),
- ("auto_keywords_float_not_allowed", {"auto_keywords": 3.14}, "Input should be a valid integer"),
- ("auto_keywords_type_invalid", {"auto_keywords": "string"}, "Input should be a valid integer"),
- ("auto_questions_min_limit", {"auto_questions": -1}, "Input should be greater than or equal to 0"),
- ("auto_questions_max_limit", {"auto_questions": 11}, "Input should be less than or equal to 10"),
- ("auto_questions_float_not_allowed", {"auto_questions": 3.14}, "Input should be a valid integer"),
- ("auto_questions_type_invalid", {"auto_questions": "string"}, "Input should be a valid integer"),
- ("chunk_token_num_min_limit", {"chunk_token_num": 0}, "Input should be greater than or equal to 1"),
- ("chunk_token_num_max_limit", {"chunk_token_num": 2049}, "Input should be less than or equal to 2048"),
- ("chunk_token_num_float_not_allowed", {"chunk_token_num": 3.14}, "Input should be a valid integer"),
- ("chunk_token_num_type_invalid", {"chunk_token_num": "string"}, "Input should be a valid integer"),
- ("delimiter_empty", {"delimiter": ""}, "String should have at least 1 character"),
- ("html4excel_type_invalid", {"html4excel": "string"}, "Input should be a valid boolean"),
- ("tag_kb_ids_not_list", {"tag_kb_ids": "1,2"}, "Input should be a valid list"),
- ("tag_kb_ids_int_in_list", {"tag_kb_ids": [1, 2]}, "Input should be a valid string"),
- ("topn_tags_min_limit", {"topn_tags": 0}, "Input should be greater than or equal to 1"),
- ("topn_tags_max_limit", {"topn_tags": 11}, "Input should be less than or equal to 10"),
- ("topn_tags_float_not_allowed", {"topn_tags": 3.14}, "Input should be a valid integer"),
- ("topn_tags_type_invalid", {"topn_tags": "string"}, "Input should be a valid integer"),
- ("filename_embd_weight_min_limit", {"filename_embd_weight": -1}, "Input should be greater than or equal to 0"),
- ("filename_embd_weight_max_limit", {"filename_embd_weight": 1.1}, "Input should be less than or equal to 1"),
- ("filename_embd_weight_type_invalid", {"filename_embd_weight": "string"}, "Input should be a valid number"),
- ("task_page_size_min_limit", {"task_page_size": 0}, "Input should be greater than or equal to 1"),
- ("task_page_size_float_not_allowed", {"task_page_size": 3.14}, "Input should be a valid integer"),
- ("task_page_size_type_invalid", {"task_page_size": "string"}, "Input should be a valid integer"),
- ("pages_not_list", {"pages": "1,2"}, "Input should be a valid list"),
- ("pages_not_list_in_list", {"pages": ["1,2"]}, "Input should be a valid list"),
- ("pages_not_int_list", {"pages": [["string1", "string2"]]}, "Input should be a valid integer"),
- ("graphrag_type_invalid", {"graphrag": {"use_graphrag": "string"}}, "Input should be a valid boolean"),
- ("graphrag_entity_types_not_list", {"graphrag": {"entity_types": "1,2"}}, "Input should be a valid list"),
- ("graphrag_entity_types_not_str_in_list", {"graphrag": {"entity_types": [1, 2]}}, "nput should be a valid string"),
- ("graphrag_method_unknown", {"graphrag": {"method": "unknown"}}, "Input should be 'light', 'general' or 'ner'"),
- ("graphrag_method_none", {"graphrag": {"method": None}}, "Input should be 'light', 'general' or 'ner'"),
- ("graphrag_community_type_invalid", {"graphrag": {"community": "string"}}, "Input should be a valid boolean"),
- ("graphrag_resolution_type_invalid", {"graphrag": {"resolution": "string"}}, "Input should be a valid boolean"),
- ("raptor_type_invalid", {"raptor": {"use_raptor": "string"}}, "Input should be a valid boolean"),
- ("raptor_prompt_empty", {"raptor": {"prompt": ""}}, "String should have at least 1 character"),
- ("raptor_prompt_space", {"raptor": {"prompt": " "}}, "String should have at least 1 character"),
- ("raptor_max_token_max_limit", {"raptor": {"max_token": 2049}}, "Input should be less than or equal to 2048"),
- ("raptor_max_token_type_invalid", {"raptor": {"max_token": "string"}}, "Input should be a valid integer"),
- ("raptor_clustering_threshold_min_limit", {"raptor": {"clustering_threshold": -0.1}}, "Input should be greater than or equal to 0"),
- ("raptor_clustering_threshold_max_limit", {"raptor": {"clustering_threshold": 1.1}}, "Input should be less than or equal to 1"),
- ("raptor_clustering_threshold_type_invalid", {"raptor": {"clustering_threshold": "string"}}, "Input should be a valid number"),
- ("raptor_max_cluster_min_limit", {"raptor": {"max_cluster": 0}}, "Input should be greater than or equal to 1"),
- ("raptor_max_cluster_max_limit", {"raptor": {"max_cluster": 1025}}, "Input should be less than or equal to 1024"),
- ("raptor_max_cluster_float_not_allowed", {"raptor": {"max_cluster": 3.14}}, "Input should be a valid integer"),
- ("raptor_max_cluster_type_invalid", {"raptor": {"max_cluster": "string"}}, "Input should be a valid integer"),
- ("raptor_random_seed_min_limit", {"raptor": {"random_seed": -1}}, "Input should be greater than or equal to 0"),
- ("raptor_random_seed_float_not_allowed", {"raptor": {"random_seed": 3.14}}, "Input should be a valid integer"),
- ("raptor_random_seed_type_invalid", {"raptor": {"random_seed": "string"}}, "Input should be a valid integer"),
- ("parser_config_type_invalid", {"delimiter": "a" * 65536}, "Parser config exceeds size limit (max 65,535 characters)"),
- ("parent_child_type_invalid", {"parent_child": {"use_parent_child": "string"}}, "Input should be a valid boolean"),
- ("parent_child_delimiter_empty", {"parent_child": {"children_delimiter": ""}}, "String should have at least 1 character"),
- ]
- for name, parser_config, expected_message in invalid_cases:
- res = rest_client.post("/datasets", json={"name": name, "parser_config": parser_config})
- assert res.status_code == 200
- payload = res.json()
- assert payload["code"] == 101, payload
- assert expected_message in payload["message"], payload
-
-
@pytest.mark.p2
def test_dataset_create_parser_config_defaults_and_extra_fields_contract(rest_client, clear_datasets):
empty_res = rest_client.post("/datasets", json={"name": "parser_config_empty", "parser_config": {}})
@@ -1910,11 +1193,7 @@ def test_dataset_create_parser_config_defaults_and_extra_fields_contract(rest_cl
unset_parser_config = unset_payload["data"]["parser_config"]
none_parser_config = none_payload["data"]["parser_config"]
assert empty_parser_config == unset_parser_config == none_parser_config
- if IS_GO_PROXY:
- _assert_go_pipeline_parser_config(empty_parser_config)
- else:
- for key in DEFAULT_PARSER_CONFIG:
- assert key in empty_parser_config, empty_payload
+ _assert_go_pipeline_parser_config(empty_parser_config)
unsupported_field_payloads = [
{"name": "id", "id": "id"},
@@ -2410,7 +1689,7 @@ def test_dataset_metadata_config_get_and_update_contract(rest_client, create_dat
assert success_res.status_code == 200
success_payload = success_res.json()
assert success_payload["code"] == 0, success_payload
- expected_empty = {"enabled": False, "metadata": [], "built_in_metadata": []} if IS_GO_PROXY else {"metadata": [], "built_in_metadata": []}
+ expected_empty = {"enabled": False, "metadata": [], "built_in_metadata": []}
assert success_payload["data"] == expected_empty, success_payload
for scenario_name, client in (("missing token", RestClient(token=None)), ("invalid token", RestClient(token=INVALID_API_TOKEN))):
@@ -2451,11 +1730,10 @@ def test_dataset_metadata_config_get_and_update_contract(rest_client, create_dat
{"key": "size", "type": "number", "description": "File size", "enum": None},
],
}
- if IS_GO_PROXY:
- # Go keeps the dataset-level metadata.enabled flag: it is not derived
- # from whether fields are present, so a fresh dataset (flag disabled)
- # stays disabled until enabled explicitly.
- normalized_update_payload["enabled"] = False
+ # Go keeps the dataset-level metadata.enabled flag: it is not derived
+ # from whether fields are present, so a fresh dataset (flag disabled)
+ # stays disabled until enabled explicitly.
+ normalized_update_payload["enabled"] = False
update_res = rest_client.put(f"/datasets/{dataset_id}/metadata/config", json=update_payload)
assert update_res.status_code == 200
update_body = update_res.json()
@@ -2593,8 +1871,8 @@ def test_dataset_tags_and_aggregation(rest_client, create_dataset):
dataset_id = create_dataset("dataset_tags")
second_dataset_id = create_dataset("dataset_tags_second")
- # The aggregation route is served by both backends, so it is asserted
- # before the backend-specific tag list contract below.
+ # The aggregation route is asserted before the per-dataset tag routes,
+ # which the Go API does not serve.
aggregate_res = rest_client.get(
"/datasets/tags/aggregation",
params={"dataset_ids": f"{dataset_id},{second_dataset_id}"},
@@ -2608,73 +1886,24 @@ def test_dataset_tags_and_aggregation(rest_client, create_dataset):
empty_aggregate_payload = empty_aggregate_res.json()
assert empty_aggregate_payload["code"] != 0, empty_aggregate_payload
- if IS_GO_PROXY:
- # Go keeps no tag dataset, so the per-dataset tag routes are not served
- # at all; tagging is driven by the dataset's tag source file and only
- # the aggregation route survives. Assert the contract that replaced the
- # old 200 responses instead of the route's payload.
- assert rest_client.get(f"/datasets/{dataset_id}/tags").status_code == 404
- assert rest_client.get("/datasets/invalid_id/tags").status_code == 404
- return
-
- list_tags_res = rest_client.get(f"/datasets/{dataset_id}/tags")
- assert list_tags_res.status_code == 200
- list_tags_payload = list_tags_res.json()
- # Known env/runtime behavior: this route can return 102 when retriever tag
- # backend is unavailable for an empty dataset. Keep route-contract coverage.
- assert list_tags_payload["code"] in (0, 102), list_tags_payload
- if list_tags_payload["code"] == 0:
- assert isinstance(list_tags_payload["data"], list), list_tags_payload
-
- invalid_list_tags_res = rest_client.get("/datasets/invalid_id/tags")
- assert invalid_list_tags_res.status_code == 200
- invalid_list_tags_payload = invalid_list_tags_res.json()
- assert invalid_list_tags_payload["code"] != 0, invalid_list_tags_payload
+ # Go keeps no tag dataset, so the per-dataset tag routes are not served
+ # at all; tagging is driven by the dataset's tag source file and only
+ # the aggregation route survives. The 404s below replace the old 200
+ # responses.
+ assert rest_client.get(f"/datasets/{dataset_id}/tags").status_code == 404
+ assert rest_client.get("/datasets/invalid_id/tags").status_code == 404
@pytest.mark.p2
def test_dataset_tags_delete_and_rename_validation(rest_client, create_dataset):
dataset_id = create_dataset("dataset_tag_mutation")
- if IS_GO_PROXY:
- # Same removal as above: with no tag dataset there is nothing to rename
- # or delete, so both the validation cases and the invalid-dataset cases
- # are now unrouted instead of argument errors.
- for url in (f"/datasets/{dataset_id}/tags", "/datasets/invalid_id/tags"):
- assert rest_client.delete(url, json={"tags": ["tag1"]}).status_code == 404
- assert rest_client.put(url, json={"from_tag": "old", "to_tag": "new"}).status_code == 404
- return
-
- delete_missing_tags = rest_client.delete(f"/datasets/{dataset_id}/tags", json={})
- assert delete_missing_tags.status_code == 200
- delete_missing_tags_payload = delete_missing_tags.json()
- assert delete_missing_tags_payload["code"] != 0, delete_missing_tags_payload
-
- delete_invalid_tags_type = rest_client.delete(f"/datasets/{dataset_id}/tags", json={"tags": "wrong"})
- assert delete_invalid_tags_type.status_code == 200
- delete_invalid_tags_type_payload = delete_invalid_tags_type.json()
- assert delete_invalid_tags_type_payload["code"] != 0, delete_invalid_tags_type_payload
-
- delete_invalid_dataset_tags = rest_client.delete("/datasets/invalid_id/tags", json={"tags": ["tag1"]})
- assert delete_invalid_dataset_tags.status_code == 200
- delete_invalid_dataset_tags_payload = delete_invalid_dataset_tags.json()
- assert delete_invalid_dataset_tags_payload["code"] != 0, delete_invalid_dataset_tags_payload
-
- rename_empty = rest_client.put(
- f"/datasets/{dataset_id}/tags",
- json={"from_tag": "", "to_tag": ""},
- )
- assert rename_empty.status_code == 200
- rename_empty_payload = rename_empty.json()
- assert rename_empty_payload["code"] != 0, rename_empty_payload
-
- rename_invalid_dataset = rest_client.put(
- "/datasets/invalid_id/tags",
- json={"from_tag": "old", "to_tag": "new"},
- )
- assert rename_invalid_dataset.status_code == 200
- rename_invalid_dataset_payload = rename_invalid_dataset.json()
- assert rename_invalid_dataset_payload["code"] != 0, rename_invalid_dataset_payload
+ # With no tag dataset there is nothing to rename or delete, so both the
+ # validation cases and the invalid-dataset cases are unrouted instead of
+ # argument errors.
+ for url in (f"/datasets/{dataset_id}/tags", "/datasets/invalid_id/tags"):
+ assert rest_client.delete(url, json={"tags": ["tag1"]}).status_code == 404
+ assert rest_client.put(url, json={"from_tag": "old", "to_tag": "new"}).status_code == 404
@pytest.mark.p2
diff --git a/test/testcases/restful_api/test_document_raw_routes.py b/test/testcases/restful_api/test_document_raw_routes.py
index a4c7d48c66..454251c04f 100644
--- a/test/testcases/restful_api/test_document_raw_routes.py
+++ b/test/testcases/restful_api/test_document_raw_routes.py
@@ -16,7 +16,7 @@
import pytest
-from test.testcases.configs import INVALID_API_TOKEN, IS_GO_PROXY
+from test.testcases.configs import INVALID_API_TOKEN
from test.testcases.restful_api.helpers.assertions import assert_auth_error
from test.testcases.restful_api.helpers.client import RestClient
@@ -41,14 +41,10 @@ def test_document_image_invalid_id_contract(rest_client):
)
def test_private_document_images_require_auth(rest_client_noauth, path):
res = rest_client_noauth.get(path)
- if IS_GO_PROXY:
- assert res.status_code == 200, res.text
- payload = res.json()
- assert payload["code"] == 102, payload
- assert payload["message"] == "Authorization is not valid!", payload
- return
- assert res.status_code == 401, res.text
- assert_auth_error(res.json(), "missing token")
+ assert res.status_code == 200, res.text
+ payload = res.json()
+ assert payload["code"] == 102, payload
+ assert payload["message"] == "Authorization is not valid!", payload
@pytest.mark.p2
diff --git a/test/testcases/restful_api/test_documents.py b/test/testcases/restful_api/test_documents.py
index f1a6a36c1d..9d87fbbb3e 100644
--- a/test/testcases/restful_api/test_documents.py
+++ b/test/testcases/restful_api/test_documents.py
@@ -27,7 +27,7 @@ from requests_toolbelt import MultipartEncoder
from utils import wait_for
from utils.file_utils import create_txt_file
-from test.testcases.configs import DEFAULT_PARSER_CONFIG, DOCUMENT_NAME_LIMIT, HOST_ADDRESS, INVALID_API_TOKEN, INVALID_ID_32, IS_GO_PROXY, VERSION
+from test.testcases.configs import DOCUMENT_NAME_LIMIT, HOST_ADDRESS, INVALID_API_TOKEN, INVALID_ID_32, VERSION
from test.testcases.restful_api.helpers.assertions import assert_auth_error
from test.testcases.restful_api.helpers.client import RestClient
from test.testcases.utils import compare_by_hash
@@ -492,7 +492,7 @@ def test_documents_upload_error_contract(rest_client, create_dataset, tmp_path):
assert filename_empty_res.status_code == 200
filename_empty_payload = filename_empty_res.json()
assert filename_empty_payload["code"] == 101, filename_empty_payload
- expected_message = "No file part!" if IS_GO_PROXY else "No file selected!"
+ expected_message = "No file part!"
assert filename_empty_payload["message"] == expected_message, filename_empty_payload
@@ -821,28 +821,11 @@ def test_documents_update_parser_config_contract(rest_client, create_dataset, tm
list_body = list_res.json()
assert list_body["code"] == 0, (parser_config, list_body)
doc_parser_config = list_body["data"]["docs"][0]["parser_config"]
- if IS_GO_PROXY:
- assert isinstance(doc_parser_config, dict) and doc_parser_config, (parser_config, list_body)
- assert "raptor" not in doc_parser_config, (parser_config, list_body)
- assert "graphrag" not in doc_parser_config, (parser_config, list_body)
- continue
- if parser_config == {}:
- assert doc_parser_config == DEFAULT_PARSER_CONFIG, (parser_config, list_body)
- else:
- for key, value in parser_config.items():
- if key in {"graphrag", "raptor"}:
- assert key not in doc_parser_config, (parser_config, list_body)
- continue
- if isinstance(value, dict):
- for sub_key, sub_value in value.items():
- assert doc_parser_config[key][sub_key] == sub_value, (parser_config, list_body)
- else:
- assert doc_parser_config[key] == value, (parser_config, list_body)
+ assert isinstance(doc_parser_config, dict) and doc_parser_config, (parser_config, list_body)
+ assert "raptor" not in doc_parser_config, (parser_config, list_body)
+ assert "graphrag" not in doc_parser_config, (parser_config, list_body)
else:
- if IS_GO_PROXY:
- assert body["message"] in expected_message, (parser_config, body)
- else:
- assert body["message"] == expected_message, (parser_config, body)
+ assert body["message"] in expected_message, (parser_config, body)
@pytest.mark.p2
@@ -1045,12 +1028,9 @@ def test_document_metadata_config_contract(rest_client, create_document):
update_body = update_res.json()
assert update_body["code"] == 0, update_body
parser_config = update_body["data"]["parser_config"]
- if IS_GO_PROXY:
- # Go scopes metadata onto the Extractor node rather than a flat key.
- scoped = parser_config.get("Extractor:AutoExtractDefault", {})
- assert scoped.get("metadata") == update_payload["metadata"], update_body
- else:
- assert parser_config["metadata"] == update_payload["metadata"], update_body
+ # Go scopes metadata onto the Extractor node rather than a flat key.
+ scoped = parser_config.get("Extractor:AutoExtractDefault", {})
+ assert scoped.get("metadata") == update_payload["metadata"], update_body
@pytest.mark.p2
diff --git a/test/testcases/restful_api/test_memories_messages.py b/test/testcases/restful_api/test_memories_messages.py
index d5acf0c8d7..69bde501f6 100644
--- a/test/testcases/restful_api/test_memories_messages.py
+++ b/test/testcases/restful_api/test_memories_messages.py
@@ -19,8 +19,6 @@ import uuid
import pytest
-from test.testcases.configs import IS_GO_PROXY
-
@pytest.fixture
def memory_cleanup(rest_client):
@@ -130,7 +128,7 @@ def test_memory_create_missing_required_fields(rest_client):
res = rest_client.post("/memories", json={"name": "missing_models", "memory_type": ["raw"]})
assert res.status_code == 200
payload = res.json()
- assert payload["code"] == (400 if IS_GO_PROXY else 101), payload
+ assert payload["code"] == 400, payload
@pytest.mark.p1
diff --git a/test/testcases/restful_api/test_openai_compatible.py b/test/testcases/restful_api/test_openai_compatible.py
index 0f1ba0497a..d411f3bba1 100644
--- a/test/testcases/restful_api/test_openai_compatible.py
+++ b/test/testcases/restful_api/test_openai_compatible.py
@@ -18,15 +18,13 @@ import json
import pytest
-from test.testcases.configs import IS_GO_PROXY
-
def _sse_events(response_text: str) -> list[str]:
return [line[5:] for line in response_text.splitlines() if line.startswith("data:")]
def skip_if_go_proxy_upstream_error(choice_message: dict) -> None:
- if IS_GO_PROXY and choice_message.get("reference") is None and choice_message.get("content", "").startswith("**ERROR**"):
+ if choice_message.get("reference") is None and choice_message.get("content", "").startswith("**ERROR**"):
pytest.skip("Go OpenAI-compatible completion could not reach the configured chat model")
@@ -96,7 +94,7 @@ def test_openai_compatible_metadata_condition_requires_object(rest_client, creat
)
assert res.status_code == 200
payload = res.json()
- assert payload["code"] == (101 if IS_GO_PROXY else 102), payload
+ assert payload["code"] == 101, payload
assert "metadata_condition must be an object." in payload["message"], payload
@@ -113,7 +111,7 @@ def test_openai_compatible_invalid_chat(rest_client):
assert res.status_code == 200
payload = res.json()
assert payload["code"] != 0, payload
- expected_message = "no authorization" if IS_GO_PROXY else "don't own the chat"
+ expected_message = "no authorization"
assert expected_message in payload["message"], payload
diff --git a/test/testcases/restful_api/test_searches.py b/test/testcases/restful_api/test_searches.py
index afed81e328..4310c1352e 100644
--- a/test/testcases/restful_api/test_searches.py
+++ b/test/testcases/restful_api/test_searches.py
@@ -18,8 +18,6 @@ import uuid
import pytest
-from test.testcases.configs import IS_GO_PROXY
-
@pytest.fixture
def search_resource(rest_client):
@@ -107,7 +105,7 @@ def test_search_update_invalid_search_id(rest_client):
@pytest.mark.p2
def test_search_completion_requires_question(rest_client, search_resource):
search_id = search_resource
- expected_message = "question is required" if IS_GO_PROXY else "required argument are missing: question"
+ expected_message = "question is required"
completion_res = rest_client.post(f"/searches/{search_id}/completion", json={})
assert completion_res.status_code == 200
diff --git a/test/testcases/restful_api/test_sessions.py b/test/testcases/restful_api/test_sessions.py
index 4feec4a7a9..1fe55bd147 100644
--- a/test/testcases/restful_api/test_sessions.py
+++ b/test/testcases/restful_api/test_sessions.py
@@ -19,7 +19,7 @@ from concurrent.futures import ThreadPoolExecutor
import pytest
-from test.testcases.configs import INVALID_API_TOKEN, INVALID_ID_32, IS_GO_PROXY, SESSION_WITH_CHAT_NAME_LIMIT
+from test.testcases.configs import INVALID_API_TOKEN, INVALID_ID_32, SESSION_WITH_CHAT_NAME_LIMIT
from test.testcases.restful_api.helpers.assertions import assert_auth_error
from test.testcases.restful_api.helpers.client import RestClient
from test.testcases.utils import is_sorted
@@ -423,7 +423,7 @@ def test_session_list_page_and_sort_contract(rest_client, create_chat):
("page negative", {"page": -1, "page_size": 2}, 0, 2, ""),
("page alpha", {"page": "a", "page_size": 2}, 0, 2, ""),
("page_size none", {"page_size": None}, 0, 5, ""),
- ("page_size zero", {"page_size": 0}, 0, 0 if IS_GO_PROXY else 5, ""),
+ ("page_size zero", {"page_size": 0}, 0, 0, ""),
("page_size one", {"page_size": 1}, 0, 1, ""),
("page_size six", {"page_size": 6}, 0, 5, ""),
("page_size negative", {"page_size": -1}, 0, 5, ""),
@@ -587,7 +587,7 @@ def test_chat_recommendation_requires_question(rest_client):
assert res.status_code == 200
payload = res.json()
assert payload["code"] == 101, payload
- expected_message = "question is required" if IS_GO_PROXY else "required argument are missing: question"
+ expected_message = "question is required"
assert expected_message in payload["message"], payload
@@ -601,16 +601,8 @@ def test_related_questions_compatibility_requires_auth(rest_client_noauth):
)
assert res.status_code == 200
payload = res.json()
- if IS_GO_PROXY:
- assert payload["code"] == 102, payload
- assert payload["message"] == "Authorization is not valid!", payload
- else:
- assert payload["code"] == 102, payload
- assert payload["message"].strip() in {
- "Authorization is not valid!",
- 'Authentication error: API key is invalid!"',
- "Authentication error: API key is invalid!",
- }, payload
+ assert payload["code"] == 102, payload
+ assert payload["message"] == "Authorization is not valid!", payload
@pytest.mark.p2