Stanislav Hejny
181 lines
6.7 KiB
Bash
181 lines
6.7 KiB
Bash
#!/bin/bash
|
|
|
|
. ./test_shared_functions.sh
|
|
|
|
JSON_FILE=./test_data/Instituto_Superior_Tecnico.md.json
|
|
HTML_FILE=./test_data/AMERICA-VENTURE_HIGHWAY.html
|
|
MP3_FILE=./test_data/flight_of_the_bumblebee_2.mp3
|
|
JPG_FILE=./test_data/China_Beijing_Summer_Palace2.jpg
|
|
EXCEL_FILE=./test_data/Worldbank-Datasheet-02082022.xlsx
|
|
TEXT_FILE=./test_data/America-Horse_with_no_name.txt
|
|
OPEN_OFFICE_FILE=./test_data/GAR_ARchitecture.odg
|
|
|
|
echo "# 1. Create Document Tests"
|
|
echo "=== Testing Document Creation ==="
|
|
|
|
echo "# Test 1.1: Create document with file, use different formats"
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${TEXT_FILE}" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_FILE=${DOCUMENT_ID}
|
|
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${JSON_FILE}" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_JSON_FILE=${DOCUMENT_ID}
|
|
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${HTML_FILE}" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_HTML_FILE=${DOCUMENT_ID}
|
|
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${MP3_FILE}" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_MP3_FILE=${DOCUMENT_ID}
|
|
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${JPG_FILE}" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_JPG_FILE=${DOCUMENT_ID}
|
|
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${EXCEL_FILE}" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_EXCEL_FILE=${DOCUMENT_ID}
|
|
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${OPEN_OFFICE_FILE}" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_OPEN_OFFICE_FILE=${DOCUMENT_ID}
|
|
|
|
echo "# Test 1.2: Create document with raw text"
|
|
make_request "POST" "/documents" \
|
|
'-F raw_text="On the first part of the journey
|
|
I was looking at all the life
|
|
There were plants and birds and rocks and things
|
|
There was sand and hills and rings
|
|
The first thing I met was a fly with a buzz
|
|
And the sky with no clouds
|
|
The heat was hot and the ground was dry
|
|
But the air was full of sound"' \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_RAW_TEXT=${DOCUMENT_ID}
|
|
|
|
echo "# Test 1.3: Create document with chunks"
|
|
make_request "POST" "/documents" \
|
|
"-F chunks=\"[\\\"I've been through the desert on a horse with no name\\\",\\\"It felt good to be out of the rain\\\"]\"" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_CHUNKS=${DOCUMENT_ID}
|
|
|
|
echo "# Test 1.4: Create document with metadata"
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${TEXT_FILE} -F metadata='{\"title\":\"Horse with No Name\",\"author\":\"America\"}'" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_FILE_WITH_METADATA=${DOCUMENT_ID}
|
|
|
|
echo "# Test 1.5: Create document with custom ingestion config"
|
|
make_request "POST" "/documents" \
|
|
"-F file=@${TEXT_FILE} -F ingestion_mode=hi-res -F ingestion_config='{\"chunk_size\":1000}'" \
|
|
"-H 'Content-Type: multipart/form-data'"
|
|
DOCUMENT_ID_FROM_FILE_WITH_CUSTOM_INGEST=${DOCUMENT_ID}
|
|
|
|
echo "# 2. List Documents Tests"
|
|
echo "=== Testing Document Listing ==="
|
|
|
|
echo "# Test 2.1: List all documents"
|
|
make_request "GET" "/documents" ""
|
|
|
|
echo "# Test 2.2: List documents with pagination"
|
|
make_request "GET" "/documents?offset=0&limit=10" ""
|
|
|
|
echo "# Test 2.3: List specific documents by IDs"
|
|
make_request "GET" "/documents?ids=${DOCUMENT_ID_FROM_CHUNKS},${DOCUMENT_ID_FROM_FILE_WITH_METADATA}" ""
|
|
|
|
echo "# Test 2.4: List documents with summary embeddings"
|
|
make_request "GET" "/documents?include_summary_embeddings=1" ""
|
|
|
|
echo "# 3. Get Document Tests"
|
|
echo "=== Testing Document Retrieval ==="
|
|
|
|
echo "# Test 3.1: Get document by ID"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_FILE_WITH_METADATA}" ""
|
|
|
|
echo "# Test 3.2: Get non-existent document"
|
|
make_request "GET" "/documents/00000000-0000-0000-0000-000000000000" ""
|
|
|
|
echo "# 4. List Document Chunks Tests"
|
|
echo "=== Testing Document Chunks ==="
|
|
|
|
echo "# Test 4.1: List chunks with pagination"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_CHUNKS}/chunks?offset=0&limit=10" ""
|
|
|
|
echo "# Test 4.2: List chunks with vectors"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_CHUNKS}/chunks?include_vectors=true" ""
|
|
|
|
echo "# 5. Download Document Tests"
|
|
echo "=== Testing Document Download ==="
|
|
|
|
echo "# Test 5.1: Download document"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_FILE}/download" ""
|
|
|
|
echo "# 6. Delete Document Tests"
|
|
echo "=== Testing Document Deletion ==="
|
|
|
|
echo "# Test 6.1: Delete document by ID"
|
|
make_request "DELETE" "/documents/${DOCUMENT_ID_FROM_RAW_TEXT}" ""
|
|
|
|
echo "# Test 6.2: Delete documents by filter"
|
|
make_request "DELETE" "/documents/by-filter" \
|
|
"-d '{\"document_type\":{\"\$eq\":\"txt\"}}'"
|
|
|
|
echo "# 7. Document Collections Tests"
|
|
echo "=== Testing Document Collections ==="
|
|
|
|
echo "# Test 7.1: List document collections"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_CHUNKS}/collections?offset=0&limit=10" ""
|
|
|
|
echo "# 8. Entity Extraction Tests"
|
|
echo "=== Testing Entity Extraction ==="
|
|
|
|
echo "# Test 8.1: Extract entities with default settings"
|
|
make_request "POST" "/documents/${DOCUMENT_ID_FROM_FILE_WITH_METADATA}/extract" \
|
|
"-d '{\"run_type\":\"RUN\"}'"
|
|
|
|
echo "# Test 8.2: Extract entities with custom settings"
|
|
make_request "POST" "/documents/${DOCUMENT_ID_FROM_FILE_WITH_METADATA}/extract" \
|
|
"-d '{\"run_type\":\"RUN\",\"settings\":{\"model\":\"gpt-4\"}}'"
|
|
|
|
echo "# 9. Entity Retrieval Tests"
|
|
echo "=== Testing Entity Retrieval ==="
|
|
|
|
echo "# Test 9.1: Get entities with pagination"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_FILE_WITH_METADATA}/entities?offset=0&limit=10" ""
|
|
|
|
echo "# Test 9.2: Get entities with embeddings"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_FILE_WITH_METADATA}/entities?include_embeddings=true" ""
|
|
|
|
echo "# 10. Relationship Tests"
|
|
echo "=== Testing Relationships ==="
|
|
|
|
echo "# Test 10.1: Get relationships with pagination"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_FILE_WITH_METADATA}/relationships?offset=0&limit=10" ""
|
|
|
|
echo "# Test 10.2: Get relationships with filters"
|
|
make_request "GET" "/documents/${DOCUMENT_ID_FROM_FILE_WITH_METADATA}/relationships?entity_names=person,company&relationship_types=works_for" ""
|
|
|
|
echo "# 11. Document Search Tests"
|
|
echo "=== Testing Document Search ==="
|
|
|
|
echo "# Test 11.1: Basic search"
|
|
make_request "POST" "/documents/search" \
|
|
"-d '{\"query\":\"test query\",\"search_mode\":\"basic\"}'"
|
|
|
|
echo "# Test 11.2: Advanced search with custom settings"
|
|
make_request "POST" "/documents/search" \
|
|
"-d '{\"query\":\"test query\",\"search_mode\":\"advanced\",\"search_settings\":{\"filters\":{\"document_type\":\"pdf\"}}}'"
|
|
|
|
echo "# Test 11.3: Custom search with complex filters"
|
|
make_request "POST" "/documents/search" \
|
|
"-d '{\"query\":\"test query\",\"search_mode\":\"custom\",\"search_settings\":{\"filters\":{\"created_at\":{\"\$gt\":\"2023-01-01\"}}}}'"
|