Get Dockerfile working

.dockerignore

@@ -0,0 +1,7 @@
+Dockerfile
+README.md
+*.pyc
+*.pyo
+*.pyd
+__pycache__
+.pytest_cache

.gitignore

@@ -1,2 +1,3 @@
+./data
 .idea
 *~

Dockerfile

@@ -27,7 +27,8 @@ FROM base as final
 
 #RUN apk add --no-cache libffi libpq
 COPY --from=builder /venv /venv
+COPY data /data
 #COPY docker-entrypoint.sh wsgi.py ./
 #CMD ["./docker-entrypoint.sh"]
 
-CMD ["/venv/bin/python", "-m", "tinysearchengine.app"]
+CMD ["/venv/bin/python", "-m", "tinysearchengine.app", "/data/index.tinysearch"]

paths.py

@@ -9,13 +9,13 @@ COMMON_CRAWL_TERMS_PATH = DATA_DIR / 'common-craw-terms.csv'
 HN_TOP_PATH = os.path.join(DATA_DIR, 'hn-top.csv')
 CRAWL_PREFIX = 'crawl_'
 CRAWL_GLOB = os.path.join(DATA_DIR, f"{CRAWL_PREFIX}*")
-INDEX_PATH = os.path.join(DATA_DIR, 'index.tinysearch')
 TEST_INDEX_PATH = os.path.join(DATA_DIR, 'index-test.tinysearch')
 TEST_TERMS_PATH = os.path.join(DATA_DIR, 'index-terms.csv')
 WIKI_DATA_PATH = os.path.join(DATA_DIR, 'enwiki-20210301-pages-articles1.xml-p1p41242.bz2')
 WIKI_TITLES_PATH = os.path.join(DATA_DIR, 'abstract-titles-sorted.txt.gz')
 
-
 DOMAINS_QUEUE_NAME = 'domains-queue-fs'
 DOMAINS_TITLES_QUEUE_NAME = 'domains-title-queue-fs'
 DOMAINS_PATH = os.path.join(DATA_DIR, 'top10milliondomains.csv.gz')
+
+INDEX_PATH = Path(__file__).parent / 'data' / 'index.tinysearch'

tinysearchengine/app.py

@@ -1,17 +1,17 @@
 import logging
+import sys
 
 import uvicorn
 
 from tinysearchengine import create_app
-
 from tinysearchengine.indexer import TinyIndex, NUM_PAGES, PAGE_SIZE, Document
-from paths import INDEX_PATH
-
-tiny_index = TinyIndex(Document, INDEX_PATH, NUM_PAGES, PAGE_SIZE)
-app = create_app.create(tiny_index)
 
 logging.basicConfig()
 
 
+index_path = sys.argv[1]
+tiny_index = TinyIndex(Document, index_path, NUM_PAGES, PAGE_SIZE)
+app = create_app.create(tiny_index)
+
 if __name__ == "__main__":
-    uvicorn.run("tinysearchengine.app:app", host="127.0.0.1", port=8080, log_level="info", reload=True)
+    uvicorn.run("tinysearchengine.app:app", host="127.0.0.1", port=8080, log_level="info")

tinysearchengine/create_app.py

@@ -1,6 +1,7 @@
 import re
 from logging import getLogger
 from operator import itemgetter
+from pathlib import Path
 
 from fastapi import FastAPI
 from starlette.responses import FileResponse
@@ -11,6 +12,7 @@ from tinysearchengine.indexer import TinyIndex, Document
 logger = getLogger(__name__)
 
 
+STATIC_FILES_PATH = Path(__file__).parent / 'static'
 SCORE_THRESHOLD = 0.25
 
 
@@ -107,7 +109,7 @@ def create(tiny_index: TinyIndex):
 
     @app.get('/')
     def index():
-        return FileResponse('tinysearchengine/static/index.html')
+        return FileResponse(STATIC_FILES_PATH / 'index.html')
 
-    app.mount('/', StaticFiles(directory="tinysearchengine/static"), name="static")
+    app.mount('/', StaticFiles(directory=STATIC_FILES_PATH), name="static")
     return app

tinysearchengine/indexer.py

@@ -2,11 +2,13 @@ import json
 import os
 from dataclasses import astuple, dataclass
 from mmap import mmap, PROT_READ
+from pathlib import Path
 from typing import TypeVar, Generic, Callable, List
 
 import mmh3
 from zstandard import ZstdDecompressor
 
+
 NUM_PAGES = 25600
 PAGE_SIZE = 4096