diff options
| author | Yuval Adam <_@yuv.al> | 2022-11-21 12:03:06 +0200 |
|---|---|---|
| committer | Yuval Adam <_@yuv.al> | 2022-11-21 12:03:06 +0200 |
| commit | c891cd7862bdb7af6bcac0c72f8561f592f6e5be (patch) | |
| tree | 1baaca966c17364282bffcf7c0366268672583f4 /scripts/scan.py | |
| parent | cf3280e780cc63d278532b46fef6b1f798b9d34e (diff) | |
Update consolidate
Diffstat (limited to 'scripts/scan.py')
| -rw-r--r-- | scripts/scan.py | 26 |
1 files changed, 20 insertions, 6 deletions
diff --git a/scripts/scan.py b/scripts/scan.py index 5741622..c83d455 100644 --- a/scripts/scan.py +++ b/scripts/scan.py @@ -106,14 +106,28 @@ class NameScanner: resps = await asyncio.gather(*tasks) def find_homepages(self): - count = 0 + res = {} for fn in listdir(DATA_DIR / "homepages"): with open(DATA_DIR / "homepages" / fn, "r") as f: - name = fn[:-5].replace(".", "") - if name in f.read(): - count += 1 - print(name) - print(count) + domain = fn[:-5] + name = domain.replace(".", "") + text = f.read() + nip = name in text + github = "github" in text + li = "linkedin" in text + sale = "for sale" in text or "register" in text or "parking" in text + res["domain"] = { + "name": name, + "name_in_page": nip, + "github": github, + "linkedin": li, + } + ge = "✅" if github else " " + le = "✅" if li else " " + ne = "✅" if nip else " " + se = "❌" if sale else " " + print(f"{name:20}{ne}{ge}{le}{se}") + print(len(res)) ns = NameScanner() |
