Fix assisted split

Add different chain for tests
Add tests infrastructure
Some refactor

.circleci/config.yml

@@ -0,0 +1,17 @@
+version: 2
+jobs:
+  build:
+    machine: true
+    steps:
+      - checkout
+      - run: docker build --tag=sbrunner/scan-to-paperless .
+      - run: docker build --tag=tests tests
+      - run: docker run --rm tests pylint process
+      - run: docker run --rm tests python3 -m pyflakes .
+      - run: docker run --rm tests bandit -r process
+      - run: docker run --rm tests mypy --ignore-missing-imports --disallow-untyped-defs --strict-optional --follow-imports skip /opt
+      - run: docker run --rm tests codespell --check-filenames --skip=./Deskew/*,*.pyc,*.png
+      - run: docker run --rm --env=PYTHONPATH=/opt/ --volume=`pwd`/results:/results tests bash -c 'mv /opt/process /opt/process.py; cd /tests; pytest --verbose --color=yes .'
+      - store_artifacts:
+          path: results
+          destination: tests-results

.dockerignore

@@ -0,0 +1,3 @@
+*
+!process
+!requirements.txt

.gitignore

@@ -0,0 +1 @@
+/results/

.pylintrc

@@ -0,0 +1 @@
+tests/.pylintrc

Dockerfile

@@ -1,5 +1,7 @@
 FROM ubuntu:cosmic
 
+COPY requirements.txt /tmp/
+
 RUN \
   . /etc/os-release && \
   apt-get update && \
@@ -10,7 +12,7 @@ RUN \
   apt-get update && \
   apt-get install --assume-yes --no-install-recommends scantailor scantailor-advanced && \
   apt-get clean && \
-  pip3 install pyyaml numpy scipy opencv-python && \
+  pip3 install --requirement=/tmp/requirements.txt && \
   rm --recursive --force /var/lib/apt/lists/* && \
   curl http://galfar.vevb.net/store/deskew-125.zip > /tmp/deskew-125.zip && \
   unzip /tmp/deskew-125.zip -d /opt && \

README.md

@@ -35,6 +35,40 @@ The scan utils will rotate and reorder all the sheets to get a good document.
 
 The options `--append-credit-card` will append all the sheets vertically to have the booth face of the credit card on the same page.
 
+### Assisted split
+
+1. Do your scan as usual with the extra option `--assisted-split`.
+
+2. After the process do his first pass you will have images with lines and numbers.
+   The lines represent the detected potential split of the image, the length indicate the strength of the detection.
+   In your config you will have somthing like:
+
+```
+assisted_split:
+- destinations:
+  - 4  # Page number of the left part of the image
+  - 1  # Same for the right page of the image
+  image: image-1.png  # name of the image
+  limits:
+  - margin: 0  # Margin around the split
+    name: 0  # Number visible on the generated image
+    value: 375  # The position of the split (can be manually edited)
+    vertical: true  # Will split the image vertically
+  - ...
+  source: /source/975468/7-assisted-split/image-1.png
+- ...
+```
+
+   Edit your config file, you should have one more destination then the limits.
+   If you put destinatination like that: 2.1, it mean that it will be the first part of the page 2 and the 2.2 will be the secound part.
+
+3. Delete the file `REMOVE_TO_CONTINUE`.
+
+4. After the process do his first pass you will have the final generated images.
+
+5. If it's OK delete the file `REMOVE_TO_CONTINUE`.
+
+
 ## Install
 
 Install in a venv in the home directory:

post_db_create.sql

@@ -0,0 +1,20 @@
+ALTER TABLE documents_correspondent ALTER COLUMN name TYPE text;
+ALTER TABLE documents_correspondent ALTER COLUMN slug TYPE text;
+ALTER TABLE documents_correspondent ALTER COLUMN match TYPE text;
+ALTER TABLE documents_tag ALTER COLUMN name TYPE text;
+ALTER TABLE documents_tag ALTER COLUMN slug TYPE text;
+ALTER TABLE documents_tag ALTER COLUMN match TYPE text;
+ALTER TABLE documents_document ALTER COLUMN title TYPE text;
+ALTER TABLE documents_document ALTER COLUMN title TYPE text;
+
+DROP INDEX documents_document_content_aa150741;
+DROP INDEX documents_document_content_aa150741_like;
+
+ALTER TABLE django_admin_log ALTER COLUMN object_repr TYPE text;
+
+CREATE EXTENSION pg_trgm;
+DROP INDEX documents_document_content;
+CREATE INDEX documents_document_content ON documents_document USING GIN (upper(content) gin_trgm_ops);
+CREATE INDEX documents_correspondent_name ON documents_correspondent USING GIN (upper(name) gin_trgm_ops);
+CREATE INDEX documents_document_title ON documents_document USING GIN (upper(title) gin_trgm_ops);
+CREATE INDEX documents_tag_name ON documents_tag USING GIN (upper(name) gin_trgm_ops);