py-code commited on
Commit
9d81d70
·
verified ·
1 Parent(s): c4574c1

Add the 60,000 most downloaded projects (354,763 projects)

Browse files
Files changed (5) hide show
  1. README.md +12 -10
  2. own_modules.parquet +2 -2
  3. projects.parquet +2 -2
  4. pth_files.parquet +2 -2
  5. refs.parquet +2 -2
README.md CHANGED
@@ -24,8 +24,9 @@ configs:
24
 
25
  # py-code.org: which PyPI packages use which names
26
 
27
- For the latest release of 335,146 PyPI projects released between 13 Nov 2025 and 13 Nov 2026,
28
- these tables list every module, function and class the code imports or reaches by attribute,
 
29
  such as `requests.Session` or `numpy.linalg.norm`. This is the data behind the search on
30
  [py-code.org](https://py-code.org). The scanner that built it is
31
  [pypi-data/py-code](https://github.com/pypi-data/py-code) (MIT).
@@ -34,7 +35,7 @@ The tables hold names, counts and file paths. They hold no source code.
34
 
35
  ## Tables
36
 
37
- ### `refs` (38.3 million rows, 763 MB)
38
 
39
  One row per project and name. Sorted by `name`, then `project`.
40
 
@@ -50,7 +51,7 @@ A name is recorded with every prefix: `os.path.join` also adds `os` and `os.path
50
  chains are followed up to 4 levels after the imported name. Names from the project's own
51
  top-level modules are left out, so a package does not count as a user of itself.
52
 
53
- ### `projects` (335,146 rows, 20 MB)
54
 
55
  | Column | Type | Meaning |
56
  |---|---|---|
@@ -61,23 +62,24 @@ top-level modules are left out, so a package does not count as a user of itself.
61
  | `summary` | string | PyPI summary, cut at 300 characters |
62
  | `file` | string | File that was scanned |
63
  | `file_size` | int | Size of that file in bytes |
 
64
  | `py_files` | int | `.py` files in the file |
65
  | `py_failed` | int | `.py` files that do not parse as current Python (mostly Python 2) |
66
  | `downloads_30d` | int | Downloads in the 30 days before the scan, for the top 60,000 projects; 0 for the rest |
67
 
68
- ### `own_modules` (382,127 rows)
69
 
70
  Top-level modules each project provides, from its file layout. Use it to go from an import
71
  name to a project: `PIL` to `pillow`, `yaml` to `PyYAML`.
72
 
73
- ### `pth_files` (626 rows)
74
 
75
  `.pth` files that contain an `import` line. Python runs these at every start-up.
76
 
77
  ## How it was built
78
 
79
  1. List every project whose latest event is newer than the cut-off date (PyPI Simple API and
80
- XML-RPC changelog).
81
  2. For each one, take the latest release and pick one file: a pure-Python wheel, then a
82
  platform wheel, then the source distribution. Wheels are read with HTTP range requests,
83
  so only the `.py` and `.pth` files are downloaded.
@@ -85,13 +87,13 @@ name to a project: `PIL` to `pillow`, `yaml` to `PyYAML`.
85
  on imported names.
86
  4. Add 30-day download counts from ClickHouse's public PyPI dataset.
87
 
88
- 12.3 million `.py` files were parsed. The scanner's README has the details and the measured run
89
- times.
90
 
91
  ## Known limits
92
 
93
  - Only the latest release of each project. Projects with no release in the window are not
94
- included.
95
  - Dynamic imports (`importlib.import_module`, `__import__`, `getattr`) are missed.
96
  - Names reached through variables (`x = np; x.foo`) are missed.
97
  - Files that do not parse as current Python are counted in `py_failed` but not indexed.
 
24
 
25
  # py-code.org: which PyPI packages use which names
26
 
27
+ For the latest release of 354,763 PyPI projects (every project released between 13 Nov 2025 and
28
+ 13 Nov 2026, plus the most downloaded projects even when their latest release is older), these
29
+ tables list every module, function and class the code imports or reaches by attribute,
30
  such as `requests.Session` or `numpy.linalg.norm`. This is the data behind the search on
31
  [py-code.org](https://py-code.org). The scanner that built it is
32
  [pypi-data/py-code](https://github.com/pypi-data/py-code) (MIT).
 
35
 
36
  ## Tables
37
 
38
+ ### `refs` (39.9 million rows, 792 MB)
39
 
40
  One row per project and name. Sorted by `name`, then `project`.
41
 
 
51
  chains are followed up to 4 levels after the imported name. Names from the project's own
52
  top-level modules are left out, so a package does not count as a user of itself.
53
 
54
+ ### `projects` (354,763 rows, 22 MB)
55
 
56
  | Column | Type | Meaning |
57
  |---|---|---|
 
62
  | `summary` | string | PyPI summary, cut at 300 characters |
63
  | `file` | string | File that was scanned |
64
  | `file_size` | int | Size of that file in bytes |
65
+ | `bytes_fetched` | int | Bytes the scanner downloaded to read the `.py` files |
66
  | `py_files` | int | `.py` files in the file |
67
  | `py_failed` | int | `.py` files that do not parse as current Python (mostly Python 2) |
68
  | `downloads_30d` | int | Downloads in the 30 days before the scan, for the top 60,000 projects; 0 for the rest |
69
 
70
+ ### `own_modules` (402,861 rows)
71
 
72
  Top-level modules each project provides, from its file layout. Use it to go from an import
73
  name to a project: `PIL` to `pillow`, `yaml` to `PyYAML`.
74
 
75
+ ### `pth_files` (748 rows)
76
 
77
  `.pth` files that contain an `import` line. Python runs these at every start-up.
78
 
79
  ## How it was built
80
 
81
  1. List every project whose latest event is newer than the cut-off date (PyPI Simple API and
82
+ XML-RPC changelog), plus the 60,000 most downloaded projects.
83
  2. For each one, take the latest release and pick one file: a pure-Python wheel, then a
84
  platform wheel, then the source distribution. Wheels are read with HTTP range requests,
85
  so only the `.py` and `.pth` files are downloaded.
 
87
  on imported names.
88
  4. Add 30-day download counts from ClickHouse's public PyPI dataset.
89
 
90
+ 12.9 million `.py` files were parsed. The index was last rebuilt on 13 Nov 2026. The scanner's README
91
+ has the details and the measured run times.
92
 
93
  ## Known limits
94
 
95
  - Only the latest release of each project. Projects with no release in the window are not
96
+ included unless they are among the 60,000 most downloaded.
97
  - Dynamic imports (`importlib.import_module`, `__import__`, `getattr`) are missed.
98
  - Names reached through variables (`x = np; x.foo`) are missed.
99
  - Files that do not parse as current Python are counted in `py_failed` but not indexed.
own_modules.parquet CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ea953eedfdcece9e0971744b78c4e3f91747ec4e495397a41a28532d233480c7
3
- size 4244864
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e08a510bf6b9feeb054d60d28d5f35bc7527d130935c19681b150428920db2ba
3
+ size 4456666
projects.parquet CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1074ce5ab96d725a52820b307b559df6b500268ec1324b49a8e688d2f4eed84a
3
- size 19530196
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72ed44601ba165f98ef05abd974de645c7c6f5fb0bb19f6042bbab7bdf244dc3
3
+ size 21752971
pth_files.parquet CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7374c4bf85e6832b4e4348a2bdabdb5a0f6af71a54774b53c2bab2f764c1e00a
3
- size 14386
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4ce64115f2e6e96343cf3c3d170936275d72d994f3b6e4e2f5aa8110964eb78
3
+ size 17237
refs.parquet CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5717aa0cf94ed8b53dfd67c72ec0160fb3ba3773b1e35973515842c4e24488e2
3
- size 763414221
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:81d3006e96c7dc92c7daaa2903a1043401817b4a9260b8fb3aed22dd58d41323
3
+ size 792364858