Merge pull request #17 from fabianhe/dev

Add JSONL output and stdin commits

.pre-commit-config.yaml

@@ -25,3 +25,11 @@ repos:
     rev: v0.812
     hooks:
       - id: mypy
+  - repo: local
+    hooks:
+      - id: typerdocs
+        name: Make typer docs
+        language: system
+        pass_filenames: false
+        always_run: true
+        entry: poetry run typer pyrepositoryminer.main utils docs --name pyrepositoryminer --output DOCS.md

DOCS.md

@@ -0,0 +1,104 @@
+# `pyrepositoryminer`
+
+Efficient Repository Mining in Python.
+
+**Usage**:
+
+```console
+$ pyrepositoryminer [OPTIONS] COMMAND [ARGS]...
+```
+
+**Options**:
+
+* `--install-completion`: Install completion for the current shell.
+* `--show-completion`: Show completion for the current shell, to copy it or customize the installation.
+* `--help`: Show this message and exit.
+
+**Commands**:
+
+* `analyze`: Analyze commits of a repository.
+* `branch`: Get the branches of a repository.
+* `clone`: Clone a repository to a path.
+* `commits`: Get the commit ids of a repository.
+
+## `pyrepositoryminer analyze`
+
+Analyze commits of a repository.
+
+**Usage**:
+
+```console
+$ pyrepositoryminer analyze [OPTIONS] REPOSITORY [METRICS]:[complexity|filecount|halstead|halstead_total|linelength|loc|maintainability|nesting|raw]...
+```
+
+**Arguments**:
+
+* `REPOSITORY`: [required]
+* `[METRICS]:[complexity|filecount|halstead|halstead_total|linelength|loc|maintainability|nesting|raw]...`
+
+**Options**:
+
+* `--commits FILENAME`
+* `--help`: Show this message and exit.
+
+## `pyrepositoryminer branch`
+
+Get the branches of a repository.
+
+**Usage**:
+
+```console
+$ pyrepositoryminer branch [OPTIONS] PATH
+```
+
+**Arguments**:
+
+* `PATH`: [required]
+
+**Options**:
+
+* `--local / --no-local`: [default: True]
+* `--remote / --no-remote`: [default: True]
+* `--help`: Show this message and exit.
+
+## `pyrepositoryminer clone`
+
+Clone a repository to a path.
+
+**Usage**:
+
+```console
+$ pyrepositoryminer clone [OPTIONS] URL PATH
+```
+
+**Arguments**:
+
+* `URL`: [required]
+* `PATH`: [required]
+
+**Options**:
+
+* `--help`: Show this message and exit.
+
+## `pyrepositoryminer commits`
+
+Get the commit ids of a repository.
+
+**Usage**:
+
+```console
+$ pyrepositoryminer commits [OPTIONS] REPOSITORY
+```
+
+**Arguments**:
+
+* `REPOSITORY`: [required]
+
+**Options**:
+
+* `--branches FILENAME`
+* `--simplify-first-parent / --no-simplify-first-parent`: [default: True]
+* `--drop-duplicates / --no-drop-duplicates`: [default: False]
+* `--sort [topological|time]`
+* `--sort-reverse / --no-sort-reverse`: [default: False]
+* `--help`: Show this message and exit.

README.md

@@ -1,6 +1,7 @@
 # pyrepositoryminer
 
 [![CI workflow](https://github.com/fabianhe/pyrepositoryminer/actions/workflows/test.yaml/badge.svg)](https://github.com/fabianhe/pyrepositoryminer/actions/workflows/test.yaml)
+![PyPI](https://img.shields.io/pypi/v/pyrepositoryminer?color=%23000)
 [![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
 
 Efficient Repository Mining in Python

pyproject.toml

@@ -1,13 +1,13 @@
 [tool.poetry]
 name = "pyrepositoryminer"
-version = "0.1.0"
+version = "0.2.0"
 description = "Efficient Repository Mining in Python"
 license = "GPL-3.0-or-later"
 authors = ["Fabian Heseding <[email protected]>"]
 readme = "README.md"
 homepage = "https://github.com/fabianhe/pyrepositoryminer"
 repository = "https://github.com/fabianhe/pyrepositoryminer"
-documentation = "https://github.com/fabianhe/pyrepositoryminer"
+documentation = "https://github.com/fabianhe/pyrepositoryminer/blob/master/DOCS.md"
 
 [tool.poetry.scripts]
 pyrepositoryminer = "pyrepositoryminer.main:app"
@@ -25,6 +25,7 @@ mypy = "^0.812"
 flake8 = "^3.9.1"
 isort = "^5.8.0"
 pre-commit = "^2.12.1"
+typer-cli = "^0.0.11"
 
 [build-system]
 requires = ["poetry-core>=1.0.0"]

pyrepositoryminer/__init__.py

@@ -1 +1 @@
-__version__ = "0.1.0"
+__version__ = "0.2.0"

pyrepositoryminer/helper.py

@@ -0,0 +1,74 @@
+from json import dumps
+from typing import Any, Iterable, List, TypedDict
+
+from pygit2 import Commit, Signature
+
+
+class Metric(TypedDict):
+    name: str
+    value: Any
+
+
+class ObjectOutput(TypedDict):
+    id: str
+
+
+class UnitOutput(ObjectOutput):
+    metrics: List[Metric]
+
+
+class BlobOutput(ObjectOutput):
+    metrics: List[Metric]
+    units: List[UnitOutput]
+
+
+class SignatureOutput(TypedDict):
+    email: str
+    name: str
+    time_offset: int
+    time: int
+
+
+class CommitBase(ObjectOutput):
+    author: SignatureOutput
+    commit_time: int
+    commit_time_offset: int
+    committer: SignatureOutput
+    message: str
+    parent_ids: List[str]
+
+
+class CommitOutput(CommitBase, total=False):
+    metrics: List[Metric]
+    blobs: List[BlobOutput]
+
+
+def parse_signature(signature: Signature) -> SignatureOutput:
+    return SignatureOutput(
+        email=str(signature.email),
+        name=str(signature.name),
+        time_offset=int(signature.offset),
+        time=int(signature.time),
+    )
+
+
+def parse_commit(
+    commit: Commit,
+    metrics: Iterable[Metric],
+    blobs: Iterable[BlobOutput],
+) -> CommitOutput:
+    return CommitOutput(
+        id=str(commit.id),
+        author=parse_signature(commit.author),
+        commit_time=int(commit.commit_time),
+        commit_time_offset=int(commit.commit_time_offset),
+        committer=parse_signature(commit.committer),
+        message=str(commit.message),
+        parent_ids=[str(id) for id in commit.parent_ids],
+        metrics=list(metrics),
+        blobs=list(blobs),
+    )
+
+
+def format_output(output: CommitOutput) -> str:
+    return dumps(output, separators=(",", ":"), indent=None)