20211018

体质检测管理.ipynb

@@ -69,27 +69,12 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 77,
+   "execution_count": null,
    "id": "91a8f4f4-c417-4db7-94ba-2d3404e5c2c9",
    "metadata": {
-    "execution": {
-     "iopub.execute_input": "2021-10-18T02:34:07.188682Z",
-     "iopub.status.busy": "2021-10-18T02:34:07.187761Z",
-     "iopub.status.idle": "2021-10-18T02:34:07.230533Z",
-     "shell.execute_reply": "2021-10-18T02:34:07.227725Z",
-     "shell.execute_reply.started": "2021-10-18T02:34:07.188577Z"
-    },
     "tags": []
    },
-   "outputs": [
+   "outputs": [],
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "ok!\n"
-     ]
-    }
-   ],
    "source": [
     "import pymongo\n",
     "from bson.objectid import ObjectId\n",
@@ -179,27 +164,12 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 79,
+   "execution_count": null,
    "id": "82e1b982-3c88-4f01-ad94-49177c4a46f6",
    "metadata": {
-    "execution": {
-     "iopub.execute_input": "2021-10-18T02:36:32.491973Z",
-     "iopub.status.busy": "2021-10-18T02:36:32.491036Z",
-     "iopub.status.idle": "2021-10-18T02:36:34.395918Z",
-     "shell.execute_reply": "2021-10-18T02:36:34.393177Z",
-     "shell.execute_reply.started": "2021-10-18T02:36:32.491866Z"
-    },
     "tags": []
    },
-   "outputs": [
+   "outputs": [],
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "<pymongo.results.InsertManyResult object at 0x7f6f28531ec0>\n"
-     ]
-    }
-   ],
    "source": [
     "import json\n",
     "import pymongo\n",
@@ -250,7 +220,6 @@
     "        dict3['sn'] = k\n",
     "        dict3['result'] = res[k]\n",
     "        dict2['exam'].append(dict3)\n",
-    "\n",
     "        list1.append(dict2)\n",
     "x = mycol.insert_many(list1)\n",
     "mycol1 = mydb[\"event\"]\n",
@@ -303,14 +272,6 @@
     "    "
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "44149ba0-1730-43d4-b07f-82a53dca4ce4",
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  },
   {
    "cell_type": "markdown",
    "id": "2bb3cbdd-9fe1-4345-8c5a-53acb1a386de",
@@ -321,29 +282,12 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 75,
+   "execution_count": null,
    "id": "910c3fcd-121a-42c0-9465-ad79a5829239",
    "metadata": {
-    "execution": {
-     "iopub.execute_input": "2021-10-18T02:29:20.586805Z",
-     "iopub.status.busy": "2021-10-18T02:29:20.585860Z",
-     "iopub.status.idle": "2021-10-18T02:29:24.051120Z",
-     "shell.execute_reply": "2021-10-18T02:29:24.049117Z",
-     "shell.execute_reply.started": "2021-10-18T02:29:20.586699Z"
-    },
     "tags": []
    },
-   "outputs": [
+   "outputs": [],
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "\n",
-      "运动项目信息：\n",
-      "ok\n"
-     ]
-    }
-   ],
    "source": [
     "import pymysql\n",
     "import json\n",
@@ -542,10 +486,62 @@
     "print(item['1'])"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "c47e9547-d896-4fe2-8f70-8635c92fa792",
+   "metadata": {},
+   "source": [
+    "## 压缩体检报告"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
    "id": "1ccab0f2-448b-4423-ae21-aa1c0f6470da",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from pdf2image import convert_from_path, convert_from_bytes\n",
+    "import img2pdf  \n",
+    "import glob\n",
+    "import os,sys\n",
+    "import tempfile\n",
+    "from pdf2image.exceptions import (\n",
+    "    PDFInfoNotInstalledError,\n",
+    "    PDFPageCountError,\n",
+    "    PDFSyntaxError\n",
+    ")\n",
+    "#images = convert_from_path('1.pdf',dpi=200,fmt='jpg', output_folder='./sample_data')\n",
+    "fl_path = 'file/210924'\n",
+    "new_path ='file/new'\n",
+    "old = glob.glob(f'{fl_path}/*.pdf')\n",
+    "old.sort()\n",
+    "for old_file in old:\n",
+    "    s = os.path.basename(old_file).split('.')[0].rjust(5,'0')\n",
+    "    new_file = f'{new_path}/{s}.pdf'\n",
+    "    \n",
+    "    with tempfile.TemporaryDirectory() as path:\n",
+    "        images_from_path = convert_from_path(old_file, dpi=100,fmt='jpg', output_folder=path)\n",
+    "        fl=glob.glob(f'{path}/*.jpg')\n",
+    "        fl.sort()\n",
+    "        a4inpt = (img2pdf.mm_to_pt(210),img2pdf.mm_to_pt(297))\n",
+    "        layout_fun = img2pdf.get_layout_fun(a4inpt)\n",
+    "        with open(new_file,\"wb\") as f:\n",
+    "            f.write(img2pdf.convert(fl,layout_fun=layout_fun))\n",
+    "\n",
+    "\n",
+    "    #print(path)\n",
+    "\n",
+    "    print(new_file,'ok')\n",
+    "#print(old)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5af86665-8abc-493d-a47f-fd9c53b8caea",
    "metadata": {},
    "outputs": [],
    "source": []