Feature 152 folder store

.github/workflows/poetry-package-test.yml

@@ -16,7 +16,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: ["3.9", "3.10"]
+        python-version: ["3.9", "3.10", "3.11"]
 
     steps:
     - uses: actions/checkout@v3

.github/workflows/uv-package-test.yml

@@ -0,0 +1,28 @@
+name: Python CI with uv and pytest
+on:
+  workflow_dispatch:
+  push:
+    branches: [ "main"]
+  pull_request:
+    branches: [ "main" ]
+jobs:
+  build:  
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.9", "3.10", "3.11", "3.12", "3.13", "3.14" ]
+    steps:
+    - uses: actions/checkout@v5
+
+    - name: Setup uv
+      uses: astral-sh/setup-uv@v7
+      with:
+        version: '0.9.15'
+        python-version: ${{ matrix.python-version }}
+
+    - name: Install the project
+      run: uv sync --all-extras --dev
+
+    - name: Run tests with pytest
+      run: uv run pytest tests

README.md

@@ -316,13 +316,23 @@ use_multiprocessing = os.getenv("USE_MULTIPROCESSING", "False") == "True"
 
 ## Development build
 
-HashStore is a python package, and built using the [Python Poetry](https://python-poetry.org)
-build tool.
-
-To install `hashstore` locally, create a virtual environment for python 3.9+,
-install poetry, and then install or build the package with `poetry install` or `poetry build`,
-respectively. Note, installing `hashstore` with poetry will also make the `hashstore` command 
-available through the command line terminal (see `HashStore Client` section below for details).
+HashStore is a python package. We recommend installing it using `uv`. Instructions on how to install and set up `uv` can be found [here](https://gist.github.com/datadavev/3975f244e5db500ba0328ef771ca74dd).
+
+Friendly Notes:
+ - You may run into a `command not found: compdef` when adding code to your `.zshrc` file, this can be resolved by adjusting the code to be:
+    ```sh
+    # .zshrc
+    autoload -Uz compinit
+    compinit
+    eval "$(uv generate-shell-completion zsh)"
+    eval "$(uvx --generate-shell-completion zsh)"
+    ```
+  - When downloading the script `uv-python-symlink`, an extension may be added to it, for example: `uv-python-symlink.txt`. It may also not have an executable status. You can execute the following to adjust it:
+    ```sh
+    $ mv uv-python-symlink uv-python-symlink.sh
+    chmod +x uv-python-symlink.sh
+    ```
+  - After following the steps and navigating to the python project, `uv` may not have sufficient permissions to run. Follow the given prompts and execute `direnv allow`
 
 To run tests, navigate to the root directory and run `pytest`. The test suite contains tests that
 take a longer time to run (relating to the storage of large files) - to execute all tests, run

folder_operations.md

@@ -0,0 +1,114 @@
+# Folders in HashStore
+
+Describes storing directory trees in hashstore.
+
+## Assumptions
+
+- The root of a folder hierarchy is identified by a PID
+- A folder hierarchy (including content) identified by a PID is immutable
+- A mutation to a folder hierarchy results in a new folder hierarchy identified by a new PID
+- Any subfolder may optionally be identified by a PID
+- Any file contained within a folder hierarchy may be identified by a PID
+- Permissions are associated with a PID and so apply to content of PID identified containers or files.
+- A folder hierarchy may reference all or part of another identified folder hierarchy
+- A folder is represented by a `container` in hashstore.
+
+## Containers
+
+Hashstore is augmented by adding an additional type of content that represents a `container`, the contents of which represent a single folder. A `container` has two types of entries: `file` that represents a single file and `folder` which represents a single subfolder. Each entry in a `container` has properties: `type`, `cid`, and `name`, where:
+
+`type` - Indicates if the entry is a folder (`0`) or file (`1`).
+
+`cid` - The content ID for the respective file or container.
+
+`name` - The name component of the path to the entry. i.e. The last path segment for a subfolder or the file name (without path) for a file.
+
+The CID for a container is computed from the serialized content on the container which includes the CID values for any subfolders. Hence, computing the CID for folders in a hierarchy requires a depth-first approach where the CIDs for leaves of a branch are computed before the branch.
+
+A container is serialized space delimited rows in a text file. Each row represents an entry in the container, with values `type`, `cid`, and `name` in that order. Since folder or file names *may* contain whitespace, the `name` entry consumes the remainder of the row.
+
+Since the CID for a container is dependent on its content, the content order is sorted by the `type` and `cid` values so hashing is consistent. Hence rows referencing subfolder containers will always appear before rows referencing files.
+
+For example, given the folder hierarchy:
+
+```
+PID_1             <- dbc15
+├── A             <- ad5eb
+│   ├── a1.txt
+│   └── a2.txt
+└── B             <- cc08d
+    └── b1.csv
+```
+
+The following `container` entries are created (`cid` values are truncated):
+
+Container `ad5eb`:
+```
+1 10fbd a1.txt
+1 c880c a2.txt
+```
+
+Container `cc08d`:
+```
+1 00e99 b1.csv
+```
+
+Container `dbc15`:
+```
+0 ad5eb A
+0 cc08d B
+```
+
+The hashstore entry for `PID_1` might be:
+```
+$ cat refs/pids/53/b2/f2/58a2f3061a7bee4ba8b157aab217795c4692e2a2d8856e2fd97eb7fa3f
+dbc1516e49e7437ea441f279570d32b1e2f149c44ab0a77682629215f4a5970b
+
+$ cat refs/cids/db/c1/51/6e49e7437ea441f279570d32b1e2f149c44ab0a77682629215f4a5970b
+PID_1
+```
+
+Each container is resolveable by the combination of PID and path. So for example,
+the folder `B` within the context of `PID_1` can be resolved using the identifier `PID_1 B`. 
+Similarly, the file `A/a2.txt` can be resolved with the identifier `PID_1 A/a2.txt`. 
+Corresponding entries in hashstore `refs/pids` and `refs/cids` are created.
+
+## Operations
+
+### Get an object by path
+
+Given a PID and a path, retrieve the corresponding object (file or folder) from hashstore.
+
+Persistent identifiers for objects within a folder hierarchy are constructed by concatenating the PID with the path using a space as a delimiter. For example, to retrieve the object at path `data/file1.txt` within the folder hierarchy identified by PID `abc123`, the identifier would be `abc123 data/file1.txt`.
+
+```
+hashstore = HashStore(...)
+path_pid = "<PID>" + " " + "<path>"
+object_stream = hashstore.retrieve_object(path_pid)
+```
+
+### Store a new folder hierarchy
+
+To store a new folder hierarchy, recursively create `container` entries for each folder in the hierarchy, starting from the leaves and working up to the root. For each folder, create a `container` with entries for its subfolders and files, compute the CID for the container, and store it in hashstore. Finally, associate the root container's CID with the PID representing the entire folder hierarchy.
+
+This is achieved by the `hashstore.store_folder()` method.
+
+```
+hashstore = HashStore(...)
+pid = "<PID>"
+source_path = "<local_folder_path>"
+hashstore.store_folder(pid, source_path)
+```
+
+### Retrieve folder hierarchy structure
+
+To retrieve the structure of a folder hierarchy identified by a PID, recursively resolve each `container` starting from the root PID. For each folder, read its `container` entries to identify subfolders and files, and continue resolving subfolders until the entire hierarchy is reconstructed.
+
+This is achieved by the `hashstore.retrieve_folder()` method.
+
+```
+hashstore = HashStore(...)
+pid = "<PID>"
+destination_path = "<local_folder_path>"
+hashstore.retrieve_folder(pid, destination_path)
+```